Milestone 3 - Deep Learning.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"provenance":[{"file_id":"1Ttqn9dEkP0b6iNMdDUNjd_sIQRsqbh8W","timestamp":1731901699577}],"gpuType":"T4","collapsed_sections":["uFr2KFpKqam_"],"toc_visible":true},"kernelspec":{"name":"python3","display_name":"Python 3"},"language_info":{"name":"python"},"accelerator":"GPU"},"cells":[{"cell_type":"markdown","source":["# Milestone 3 - Deep Learning"],"metadata":{"id":"Ri4uR41ga3E9"}},{"cell_type":"markdown","source":["Aman Gupta, Aryan Bhardwaj, Liza George, Justin Kim"],"metadata":{"id":"3kR3VOuGJLyz"}},{"cell_type":"markdown","source":["# Final Model\n","\n","In this stage we used our learnings from the data exploration last week to build a deep learning model that predicts end station based on the following input features:\n","- ride type (classic vs electric bike)\n","- rider type (member vs casual rider)\n","- start region\n","- ride start hour of the day\n","- ride start day of the month\n","- ride start month of the year\n","- ride start day of the week\n","- weather of the day\n","\n","We decided to examine the ride's start and end region (NYC neighborhood) instead of exact station becayse there were too many unique stations (474) to use directly in predictions.\n","\n","The model is a fusion of a standard neural network and a graph neural network, where start and end station regions where the nodes in the graph."],"metadata":{"id":"ug2s1shqY643"}},{"cell_type":"markdown","source":["## Import and preprocess data"],"metadata":{"id":"oEIGcZixwIEw"}},{"cell_type":"markdown","source":["### Import Relevant Packages"],"metadata":{"id":"are_kpPTZvNv"}},{"cell_type":"code","source":["!pip install torch_geometric"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"FRAfk2aPa08T","executionInfo":{"status":"ok","timestamp":1731907939178,"user_tz":360,"elapsed":4869,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"392bfac7-76c1-4098-d314-466e089b43e7"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Collecting torch_geometric\n","  Downloading torch_geometric-2.6.1-py3-none-any.whl.metadata (63 kB)\n","\u001b[?25l     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m0.0/63.1 kB\u001b[0m \u001b[31m?\u001b[0m eta \u001b[36m-:--:--\u001b[0m\r\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m63.1/63.1 kB\u001b[0m \u001b[31m1.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n","\u001b[?25hRequirement already satisfied: aiohttp in /usr/local/lib/python3.10/dist-packages (from torch_geometric) (3.10.10)\n","Requirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from torch_geometric) (2024.10.0)\n","Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch_geometric) (3.1.4)\n","Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (from torch_geometric) (1.26.4)\n","Requirement already satisfied: psutil>=5.8.0 in /usr/local/lib/python3.10/dist-packages (from torch_geometric) (5.9.5)\n","Requirement already satisfied: pyparsing in /usr/local/lib/python3.10/dist-packages (from torch_geometric) (3.2.0)\n","Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from torch_geometric) (2.32.3)\n","Requirement already satisfied: tqdm in /usr/local/lib/python3.10/dist-packages (from torch_geometric) (4.66.6)\n","Requirement already satisfied: aiohappyeyeballs>=2.3.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch_geometric) (2.4.3)\n","Requirement already satisfied: aiosignal>=1.1.2 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch_geometric) (1.3.1)\n","Requirement already satisfied: attrs>=17.3.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch_geometric) (24.2.0)\n","Requirement already satisfied: frozenlist>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch_geometric) (1.5.0)\n","Requirement already satisfied: multidict<7.0,>=4.5 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch_geometric) (6.1.0)\n","Requirement already satisfied: yarl<2.0,>=1.12.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch_geometric) (1.17.1)\n","Requirement already satisfied: async-timeout<5.0,>=4.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch_geometric) (4.0.3)\n","Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch_geometric) (3.0.2)\n","Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->torch_geometric) (3.4.0)\n","Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->torch_geometric) (3.10)\n","Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->torch_geometric) (2.2.3)\n","Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->torch_geometric) (2024.8.30)\n","Requirement already satisfied: typing-extensions>=4.1.0 in /usr/local/lib/python3.10/dist-packages (from multidict<7.0,>=4.5->aiohttp->torch_geometric) (4.12.2)\n","Requirement already satisfied: propcache>=0.2.0 in /usr/local/lib/python3.10/dist-packages (from yarl<2.0,>=1.12.0->aiohttp->torch_geometric) (0.2.0)\n","Downloading torch_geometric-2.6.1-py3-none-any.whl (1.1 MB)\n","\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.1/1.1 MB\u001b[0m \u001b[31m18.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n","\u001b[?25hInstalling collected packages: torch_geometric\n","Successfully installed torch_geometric-2.6.1\n"]}]},{"cell_type":"code","source":["from google.colab import drive\n","drive.mount('/content/drive')"],"metadata":{"id":"x89QZajiLTHf","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1731907957746,"user_tz":360,"elapsed":14556,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"d6b32b47-f493-40e1-a7cf-9b408882de60"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Mounted at /content/drive\n"]}]},{"cell_type":"code","source":["import pandas as pd\n","\n","import torch\n","import torch.nn as nn\n","import torch.optim as optim\n","import torch.nn.functional as F\n","\n","import pickle\n","import math"],"metadata":{"id":"J0rlwsoGOj6H"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["from torch_geometric.data import Data\n","from torch_geometric.loader import DataLoader\n","from torch_geometric.nn import GCNConv, GATConv\n","from torch_geometric.nn import global_mean_pool\n","\n","from sklearn.preprocessing import OneHotEncoder\n","from sklearn.preprocessing import LabelEncoder\n","from sklearn.preprocessing import StandardScaler\n","from sklearn.model_selection import train_test_split"],"metadata":{"id":"WRDCsVGVSg0C"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["### Use GPU for faster computations"],"metadata":{"id":"rsckT4njZ3yL"}},{"cell_type":"code","source":["DEVICE = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n","if __name__=='__main__':\n","    print('Using device:', DEVICE)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"PjKUiyefVRRu","executionInfo":{"status":"ok","timestamp":1731908000305,"user_tz":360,"elapsed":127,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"1b11d74e-9dda-4f49-fa32-6d2300c9d60b"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Using device: cuda\n"]}]},{"cell_type":"markdown","source":["\n","### Open Pickled Data\n","\n","This is from what we extracted and pickled in Milestone 1 (updated to our [data extraction Notebook](https://colab.research.google.com/drive/1BInmWkNqitlIHXVlWWTid3fwA4y96J5j?usp=sharing))."],"metadata":{"id":"AeBneORGHwCR"}},{"cell_type":"code","source":["# NOTE FOR GRADERS - please add the folder linked below to your drive and alter this string to match the location of the \"pickled_data\" folder:\n","# https://drive.google.com/drive/folders/1upZ3Oh242b83WTBUQB4HS64M0BlY2PHE?usp=drive_link\n","shared_folder_path = f'/content/drive/MyDrive/CS547 Project Group 11/pickled_data'\n","save_path = f'{shared_folder_path}/working_dataset_2023.pkl'"],"metadata":{"id":"5IpcAfgdue7W"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["df = pd.read_pickle(save_path)\n","df = df.sample(frac=0.5).reset_index(drop=True)\n","print(df.shape)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1731908591013,"user_tz":360,"elapsed":33065,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"36cdc5b0-a173-4d8d-e779-8e95183690f7","id":"QpiMhb0Fuk9M"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["(8776757, 13)\n"]}]},{"cell_type":"code","source":["df.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":313},"executionInfo":{"status":"ok","timestamp":1731908594772,"user_tz":360,"elapsed":161,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"ae7d411b-d602-41e1-dd15-274b05ae8954","id":"0L_gGaJAuk9M"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["            ride_id  rideable_type              started_at  \\\n","0  D4A14D8E1F6B1937   classic_bike 2023-02-15 13:24:19.223   \n","1  E09948898C7E48FD  electric_bike 2023-10-18 18:35:37.087   \n","2  EFAEB18288A3CFDA   classic_bike 2023-03-09 07:48:39.009   \n","3  0968DF3B2CB020FC   classic_bike 2023-07-07 10:56:41.532   \n","4  B3357BEC0712B6A5   classic_bike 2023-09-10 11:50:00.211   \n","\n","                 ended_at            start_station_name start_station_id  \\\n","0 2023-02-15 13:28:48.001  Kosciuszko St & Nostrand Ave          4519.04   \n","1 2023-10-18 18:38:00.418            Carmine St & 6 Ave          5763.03   \n","2 2023-03-09 07:51:20.463            E 67 St & Park Ave          6932.14   \n","3 2023-07-07 11:01:08.363               W 25 St & 6 Ave          6215.04   \n","4 2023-09-10 11:54:48.737               W 13 St & 7 Ave          6030.04   \n","\n","            end_station_name end_station_id  start_lat  start_lng    end_lat  \\\n","0  Putnam Ave & Nostrand Ave        4327.09  40.690725 -73.951335  40.684020   \n","1         Carmine St & 6 Ave        5763.03  40.730386 -74.002150  40.730386   \n","2      E 59 St & Madison Ave        6801.01  40.767801 -73.965921  40.763505   \n","3            W 15 St & 6 Ave        5989.02  40.743954 -73.991449  40.738046   \n","4      Washington Pl & 6 Ave        5838.09  40.737815 -73.999947  40.732241   \n","\n","     end_lng member_casual  \n","0 -73.949770        member  \n","1 -74.002150        member  \n","2 -73.971092        casual  \n","3 -73.996430        member  \n","4 -74.000264        member  "],"text/html":["\n","  <div id=\"df-776d00fb-2541-4118-a8ab-64f46f4ff6bb\" class=\"colab-df-container\">\n","    <div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>ride_id</th>\n","      <th>rideable_type</th>\n","      <th>started_at</th>\n","      <th>ended_at</th>\n","      <th>start_station_name</th>\n","      <th>start_station_id</th>\n","      <th>end_station_name</th>\n","      <th>end_station_id</th>\n","      <th>start_lat</th>\n","      <th>start_lng</th>\n","      <th>end_lat</th>\n","      <th>end_lng</th>\n","      <th>member_casual</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>D4A14D8E1F6B1937</td>\n","      <td>classic_bike</td>\n","      <td>2023-02-15 13:24:19.223</td>\n","      <td>2023-02-15 13:28:48.001</td>\n","      <td>Kosciuszko St &amp; Nostrand Ave</td>\n","      <td>4519.04</td>\n","      <td>Putnam Ave &amp; Nostrand Ave</td>\n","      <td>4327.09</td>\n","      <td>40.690725</td>\n","      <td>-73.951335</td>\n","      <td>40.684020</td>\n","      <td>-73.949770</td>\n","      <td>member</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>E09948898C7E48FD</td>\n","      <td>electric_bike</td>\n","      <td>2023-10-18 18:35:37.087</td>\n","      <td>2023-10-18 18:38:00.418</td>\n","      <td>Carmine St &amp; 6 Ave</td>\n","      <td>5763.03</td>\n","      <td>Carmine St &amp; 6 Ave</td>\n","      <td>5763.03</td>\n","      <td>40.730386</td>\n","      <td>-74.002150</td>\n","      <td>40.730386</td>\n","      <td>-74.002150</td>\n","      <td>member</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>EFAEB18288A3CFDA</td>\n","      <td>classic_bike</td>\n","      <td>2023-03-09 07:48:39.009</td>\n","      <td>2023-03-09 07:51:20.463</td>\n","      <td>E 67 St &amp; Park Ave</td>\n","      <td>6932.14</td>\n","      <td>E 59 St &amp; Madison Ave</td>\n","      <td>6801.01</td>\n","      <td>40.767801</td>\n","      <td>-73.965921</td>\n","      <td>40.763505</td>\n","      <td>-73.971092</td>\n","      <td>casual</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>0968DF3B2CB020FC</td>\n","      <td>classic_bike</td>\n","      <td>2023-07-07 10:56:41.532</td>\n","      <td>2023-07-07 11:01:08.363</td>\n","      <td>W 25 St &amp; 6 Ave</td>\n","      <td>6215.04</td>\n","      <td>W 15 St &amp; 6 Ave</td>\n","      <td>5989.02</td>\n","      <td>40.743954</td>\n","      <td>-73.991449</td>\n","      <td>40.738046</td>\n","      <td>-73.996430</td>\n","      <td>member</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>B3357BEC0712B6A5</td>\n","      <td>classic_bike</td>\n","      <td>2023-09-10 11:50:00.211</td>\n","      <td>2023-09-10 11:54:48.737</td>\n","      <td>W 13 St &amp; 7 Ave</td>\n","      <td>6030.04</td>\n","      <td>Washington Pl &amp; 6 Ave</td>\n","      <td>5838.09</td>\n","      <td>40.737815</td>\n","      <td>-73.999947</td>\n","      <td>40.732241</td>\n","      <td>-74.000264</td>\n","      <td>member</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>\n","    <div class=\"colab-df-buttons\">\n","\n","  <div class=\"colab-df-container\">\n","    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-776d00fb-2541-4118-a8ab-64f46f4ff6bb')\"\n","            title=\"Convert this dataframe to an interactive table.\"\n","            style=\"display:none;\">\n","\n","  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n","    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n","  </svg>\n","    </button>\n","\n","  <style>\n","    .colab-df-container {\n","      display:flex;\n","      gap: 12px;\n","    }\n","\n","    .colab-df-convert {\n","      background-color: #E8F0FE;\n","      border: none;\n","      border-radius: 50%;\n","      cursor: pointer;\n","      display: none;\n","      fill: #1967D2;\n","      height: 32px;\n","      padding: 0 0 0 0;\n","      width: 32px;\n","    }\n","\n","    .colab-df-convert:hover {\n","      background-color: #E2EBFA;\n","      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n","      fill: #174EA6;\n","    }\n","\n","    .colab-df-buttons div {\n","      margin-bottom: 4px;\n","    }\n","\n","    [theme=dark] .colab-df-convert {\n","      background-color: #3B4455;\n","      fill: #D2E3FC;\n","    }\n","\n","    [theme=dark] .colab-df-convert:hover {\n","      background-color: #434B5C;\n","      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n","      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n","      fill: #FFFFFF;\n","    }\n","  </style>\n","\n","    <script>\n","      const buttonEl =\n","        document.querySelector('#df-776d00fb-2541-4118-a8ab-64f46f4ff6bb button.colab-df-convert');\n","      buttonEl.style.display =\n","        google.colab.kernel.accessAllowed ? 'block' : 'none';\n","\n","      async function convertToInteractive(key) {\n","        const element = document.querySelector('#df-776d00fb-2541-4118-a8ab-64f46f4ff6bb');\n","        const dataTable =\n","          await google.colab.kernel.invokeFunction('convertToInteractive',\n","                                                    [key], {});\n","        if (!dataTable) return;\n","\n","        const docLinkHtml = 'Like what you see? Visit the ' +\n","          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n","          + ' to learn more about interactive tables.';\n","        element.innerHTML = '';\n","        dataTable['output_type'] = 'display_data';\n","        await google.colab.output.renderOutput(dataTable, element);\n","        const docLink = document.createElement('div');\n","        docLink.innerHTML = docLinkHtml;\n","        element.appendChild(docLink);\n","      }\n","    </script>\n","  </div>\n","\n","\n","<div id=\"df-fa155b2f-4c58-49f5-a2ad-e97a04783bce\">\n","  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-fa155b2f-4c58-49f5-a2ad-e97a04783bce')\"\n","            title=\"Suggest charts\"\n","            style=\"display:none;\">\n","\n","<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n","     width=\"24px\">\n","    <g>\n","        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n","    </g>\n","</svg>\n","  </button>\n","\n","<style>\n","  .colab-df-quickchart {\n","      --bg-color: #E8F0FE;\n","      --fill-color: #1967D2;\n","      --hover-bg-color: #E2EBFA;\n","      --hover-fill-color: #174EA6;\n","      --disabled-fill-color: #AAA;\n","      --disabled-bg-color: #DDD;\n","  }\n","\n","  [theme=dark] .colab-df-quickchart {\n","      --bg-color: #3B4455;\n","      --fill-color: #D2E3FC;\n","      --hover-bg-color: #434B5C;\n","      --hover-fill-color: #FFFFFF;\n","      --disabled-bg-color: #3B4455;\n","      --disabled-fill-color: #666;\n","  }\n","\n","  .colab-df-quickchart {\n","    background-color: var(--bg-color);\n","    border: none;\n","    border-radius: 50%;\n","    cursor: pointer;\n","    display: none;\n","    fill: var(--fill-color);\n","    height: 32px;\n","    padding: 0;\n","    width: 32px;\n","  }\n","\n","  .colab-df-quickchart:hover {\n","    background-color: var(--hover-bg-color);\n","    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n","    fill: var(--button-hover-fill-color);\n","  }\n","\n","  .colab-df-quickchart-complete:disabled,\n","  .colab-df-quickchart-complete:disabled:hover {\n","    background-color: var(--disabled-bg-color);\n","    fill: var(--disabled-fill-color);\n","    box-shadow: none;\n","  }\n","\n","  .colab-df-spinner {\n","    border: 2px solid var(--fill-color);\n","    border-color: transparent;\n","    border-bottom-color: var(--fill-color);\n","    animation:\n","      spin 1s steps(1) infinite;\n","  }\n","\n","  @keyframes spin {\n","    0% {\n","      border-color: transparent;\n","      border-bottom-color: var(--fill-color);\n","      border-left-color: var(--fill-color);\n","    }\n","    20% {\n","      border-color: transparent;\n","      border-left-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","    }\n","    30% {\n","      border-color: transparent;\n","      border-left-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","      border-right-color: var(--fill-color);\n","    }\n","    40% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","    }\n","    60% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","    }\n","    80% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","      border-bottom-color: var(--fill-color);\n","    }\n","    90% {\n","      border-color: transparent;\n","      border-bottom-color: var(--fill-color);\n","    }\n","  }\n","</style>\n","\n","  <script>\n","    async function quickchart(key) {\n","      const quickchartButtonEl =\n","        document.querySelector('#' + key + ' button');\n","      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n","      quickchartButtonEl.classList.add('colab-df-spinner');\n","      try {\n","        const charts = await google.colab.kernel.invokeFunction(\n","            'suggestCharts', [key], {});\n","      } catch (error) {\n","        console.error('Error during call to suggestCharts:', error);\n","      }\n","      quickchartButtonEl.classList.remove('colab-df-spinner');\n","      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n","    }\n","    (() => {\n","      let quickchartButtonEl =\n","        document.querySelector('#df-fa155b2f-4c58-49f5-a2ad-e97a04783bce button');\n","      quickchartButtonEl.style.display =\n","        google.colab.kernel.accessAllowed ? 'block' : 'none';\n","    })();\n","  </script>\n","</div>\n","\n","    </div>\n","  </div>\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df"}},"metadata":{},"execution_count":41}]},{"cell_type":"code","source":["# Process Weather Data\n","weather_df = pd.read_csv(f'{shared_folder_path}/weather_data.csv')\n","weather_df['DATE'] = pd.to_datetime(weather_df['DATE'])\n","\n","weather_df['start_day'] = weather_df['DATE'].dt.day\n","weather_df['start_month'] = weather_df['DATE'].dt.month\n","weather_df.drop(columns=['STATION', 'NAME', 'DATE'], inplace=True)\n","weather_df.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":206},"id":"EKrKYPhvyqtG","executionInfo":{"status":"ok","timestamp":1731908597244,"user_tz":360,"elapsed":181,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"a378d1ec-3072-4ac5-e260-bbb189ab19fd"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["   PRCP  SNOW  SNWD  WESD  WESF  start_day  start_month\n","0  0.40   0.0   0.0   0.0   0.0          1            1\n","1  0.00   0.0   0.0   0.0   0.0          2            1\n","2  0.02   0.0   0.0   0.0   0.0          3            1\n","3  0.37   0.0   0.0   0.0   0.0          4            1\n","4  0.09   0.0   0.0   0.0   0.0          5            1"],"text/html":["\n","  <div id=\"df-b934dcd0-2894-4f59-aa6b-fb2b5b04c914\" class=\"colab-df-container\">\n","    <div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>PRCP</th>\n","      <th>SNOW</th>\n","      <th>SNWD</th>\n","      <th>WESD</th>\n","      <th>WESF</th>\n","      <th>start_day</th>\n","      <th>start_month</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>0.40</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>1</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>0.00</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>2</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>0.02</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>3</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>0.37</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>4</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>0.09</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>5</td>\n","      <td>1</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>\n","    <div class=\"colab-df-buttons\">\n","\n","  <div class=\"colab-df-container\">\n","    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-b934dcd0-2894-4f59-aa6b-fb2b5b04c914')\"\n","            title=\"Convert this dataframe to an interactive table.\"\n","            style=\"display:none;\">\n","\n","  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n","    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n","  </svg>\n","    </button>\n","\n","  <style>\n","    .colab-df-container {\n","      display:flex;\n","      gap: 12px;\n","    }\n","\n","    .colab-df-convert {\n","      background-color: #E8F0FE;\n","      border: none;\n","      border-radius: 50%;\n","      cursor: pointer;\n","      display: none;\n","      fill: #1967D2;\n","      height: 32px;\n","      padding: 0 0 0 0;\n","      width: 32px;\n","    }\n","\n","    .colab-df-convert:hover {\n","      background-color: #E2EBFA;\n","      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n","      fill: #174EA6;\n","    }\n","\n","    .colab-df-buttons div {\n","      margin-bottom: 4px;\n","    }\n","\n","    [theme=dark] .colab-df-convert {\n","      background-color: #3B4455;\n","      fill: #D2E3FC;\n","    }\n","\n","    [theme=dark] .colab-df-convert:hover {\n","      background-color: #434B5C;\n","      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n","      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n","      fill: #FFFFFF;\n","    }\n","  </style>\n","\n","    <script>\n","      const buttonEl =\n","        document.querySelector('#df-b934dcd0-2894-4f59-aa6b-fb2b5b04c914 button.colab-df-convert');\n","      buttonEl.style.display =\n","        google.colab.kernel.accessAllowed ? 'block' : 'none';\n","\n","      async function convertToInteractive(key) {\n","        const element = document.querySelector('#df-b934dcd0-2894-4f59-aa6b-fb2b5b04c914');\n","        const dataTable =\n","          await google.colab.kernel.invokeFunction('convertToInteractive',\n","                                                    [key], {});\n","        if (!dataTable) return;\n","\n","        const docLinkHtml = 'Like what you see? Visit the ' +\n","          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n","          + ' to learn more about interactive tables.';\n","        element.innerHTML = '';\n","        dataTable['output_type'] = 'display_data';\n","        await google.colab.output.renderOutput(dataTable, element);\n","        const docLink = document.createElement('div');\n","        docLink.innerHTML = docLinkHtml;\n","        element.appendChild(docLink);\n","      }\n","    </script>\n","  </div>\n","\n","\n","<div id=\"df-8bf6c2c3-3680-4038-940f-d350264bf2c7\">\n","  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-8bf6c2c3-3680-4038-940f-d350264bf2c7')\"\n","            title=\"Suggest charts\"\n","            style=\"display:none;\">\n","\n","<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n","     width=\"24px\">\n","    <g>\n","        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n","    </g>\n","</svg>\n","  </button>\n","\n","<style>\n","  .colab-df-quickchart {\n","      --bg-color: #E8F0FE;\n","      --fill-color: #1967D2;\n","      --hover-bg-color: #E2EBFA;\n","      --hover-fill-color: #174EA6;\n","      --disabled-fill-color: #AAA;\n","      --disabled-bg-color: #DDD;\n","  }\n","\n","  [theme=dark] .colab-df-quickchart {\n","      --bg-color: #3B4455;\n","      --fill-color: #D2E3FC;\n","      --hover-bg-color: #434B5C;\n","      --hover-fill-color: #FFFFFF;\n","      --disabled-bg-color: #3B4455;\n","      --disabled-fill-color: #666;\n","  }\n","\n","  .colab-df-quickchart {\n","    background-color: var(--bg-color);\n","    border: none;\n","    border-radius: 50%;\n","    cursor: pointer;\n","    display: none;\n","    fill: var(--fill-color);\n","    height: 32px;\n","    padding: 0;\n","    width: 32px;\n","  }\n","\n","  .colab-df-quickchart:hover {\n","    background-color: var(--hover-bg-color);\n","    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n","    fill: var(--button-hover-fill-color);\n","  }\n","\n","  .colab-df-quickchart-complete:disabled,\n","  .colab-df-quickchart-complete:disabled:hover {\n","    background-color: var(--disabled-bg-color);\n","    fill: var(--disabled-fill-color);\n","    box-shadow: none;\n","  }\n","\n","  .colab-df-spinner {\n","    border: 2px solid var(--fill-color);\n","    border-color: transparent;\n","    border-bottom-color: var(--fill-color);\n","    animation:\n","      spin 1s steps(1) infinite;\n","  }\n","\n","  @keyframes spin {\n","    0% {\n","      border-color: transparent;\n","      border-bottom-color: var(--fill-color);\n","      border-left-color: var(--fill-color);\n","    }\n","    20% {\n","      border-color: transparent;\n","      border-left-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","    }\n","    30% {\n","      border-color: transparent;\n","      border-left-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","      border-right-color: var(--fill-color);\n","    }\n","    40% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","    }\n","    60% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","    }\n","    80% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","      border-bottom-color: var(--fill-color);\n","    }\n","    90% {\n","      border-color: transparent;\n","      border-bottom-color: var(--fill-color);\n","    }\n","  }\n","</style>\n","\n","  <script>\n","    async function quickchart(key) {\n","      const quickchartButtonEl =\n","        document.querySelector('#' + key + ' button');\n","      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n","      quickchartButtonEl.classList.add('colab-df-spinner');\n","      try {\n","        const charts = await google.colab.kernel.invokeFunction(\n","            'suggestCharts', [key], {});\n","      } catch (error) {\n","        console.error('Error during call to suggestCharts:', error);\n","      }\n","      quickchartButtonEl.classList.remove('colab-df-spinner');\n","      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n","    }\n","    (() => {\n","      let quickchartButtonEl =\n","        document.querySelector('#df-8bf6c2c3-3680-4038-940f-d350264bf2c7 button');\n","      quickchartButtonEl.style.display =\n","        google.colab.kernel.accessAllowed ? 'block' : 'none';\n","    })();\n","  </script>\n","</div>\n","\n","    </div>\n","  </div>\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"weather_df","summary":"{\n  \"name\": \"weather_df\",\n  \"rows\": 365,\n  \"fields\": [\n    {\n      \"column\": \"PRCP\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.45855565876702964,\n        \"min\": 0.0,\n        \"max\": 4.38,\n        \"num_unique_values\": 70,\n        \"samples\": [\n          0.19,\n          0.4,\n          0.55\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"SNOW\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.3374082005254977,\n        \"min\": 0.0,\n        \"max\": 5.3,\n        \"num_unique_values\": 7,\n        \"samples\": [\n          0.0,\n          0.3,\n          5.3\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"SNWD\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.38646994963983694,\n        \"min\": 0.0,\n        \"max\": 5.3,\n        \"num_unique_values\": 7,\n        \"samples\": [\n          0.0,\n          0.1,\n          3.5\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"WESD\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.0,\n        \"min\": 0.0,\n        \"max\": 0.0,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          0.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"WESF\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.03851471727467218,\n        \"min\": 0.0,\n        \"max\": 0.6,\n        \"num_unique_values\": 4,\n        \"samples\": [\n          0.1\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"start_day\",\n      \"properties\": {\n        \"dtype\": \"int32\",\n        \"num_unique_values\": 31,\n        \"samples\": [\n          28\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"start_month\",\n      \"properties\": {\n        \"dtype\": \"int32\",\n        \"num_unique_values\": 12,\n        \"samples\": [\n          11\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"}},"metadata":{},"execution_count":42}]},{"cell_type":"markdown","source":["### Mapping features to integers as needed"],"metadata":{"id":"V-nxuj6hwNfI"}},{"cell_type":"code","source":["# Region coordinates derived from ChatGPT\n","\n","REGION_MAPPING = {\n","    \"Lower Manhattan\": 0,\n","    \"Midtown\": 1,\n","    \"Upper East Side\": 2,\n","    \"Upper West Side\": 3,\n","    \"East Village\": 4,\n","    \"West Village\": 5,\n","    \"Chelsea\": 6,\n","    \"Harlem\": 7,\n","    \"Downtown Brooklyn\": 8,\n","    \"North Brooklyn\": 9\n","}\n","REGION_LOCS = {\n","    0: (40.7128, -74.0060),\n","    1: (40.7549, -73.9840),\n","    2: (40.7736, -73.9566),\n","    3: (40.7870, -73.9754),\n","    4: (40.7264, -73.9818),\n","    5: (40.7347, -74.0049),\n","    6: (40.7430, -74.0018),\n","    7: (40.8116, -73.9465),\n","    8: (40.6958, -73.9936),\n","    9: (40.7081, -73.9571)\n","}"],"metadata":{"id":"nNXMsYli3hI5"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["import numpy as np\n","\n","# Convert REGION_LOCS to a NumPy array for efficient computation\n","region_ids = np.array(list(REGION_LOCS.keys()))\n","region_coords = np.array(list(REGION_LOCS.values()))\n","\n","def vectorized_haversine(lat, lon, region_coords):\n","    \"\"\"Vectorized calculation of haversine distance between points and regions.\"\"\"\n","    R = 6371  # Earth's radius in km\n","\n","    # Convert degrees to radians\n","    lat1, lon1 = np.radians(lat), np.radians(lon)\n","    region_lat, region_lon = np.radians(region_coords[:, 0]), np.radians(region_coords[:, 1])\n","\n","    # Calculate haversine formula\n","    dlat = region_lat[:, None] - lat1\n","    dlon = region_lon[:, None] - lon1\n","    a = np.sin(dlat / 2)**2 + np.cos(region_lat[:, None]) * np.cos(lat1) * np.sin(dlon / 2)**2\n","    c = 2 * np.arcsin(np.sqrt(a))\n","    return R * c  # Distance in kilometers\n","\n","# Vectorized function to find nearest region\n","def find_nearest_regions(latitudes, longitudes, region_coords, region_ids):\n","    distances = vectorized_haversine(latitudes, longitudes, region_coords)\n","    nearest_indices = np.argmin(distances, axis=0)\n","    return region_ids[nearest_indices]\n","\n","# Apply the vectorized function to the DataFrame\n","df['start_region'] = find_nearest_regions(df['start_lat'].values, df['start_lng'].values, region_coords, region_ids)\n","df['end_region'] = find_nearest_regions(df['end_lat'].values, df['end_lng'].values, region_coords, region_ids)"],"metadata":{"id":"sW56kCUYxAkP"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Encode 'rideable_type' and store the mapping\n","df['rideable_type'], rideable_type_mapping = pd.factorize(df['rideable_type'])\n","rideable_type_mapping = dict(enumerate(rideable_type_mapping))"],"metadata":{"id":"bGeSc8-6utUx"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Encode 'member_casual' and store the mapping\n","df['member_casual'], member_casual_mapping = pd.factorize(df['member_casual'])\n","member_casual_mapping = dict(enumerate(member_casual_mapping))"],"metadata":{"id":"_o-1sGl-bR-s"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Encoding 'started_at' time into different types\n","df['start_hour'] = df['started_at'].dt.hour\n","df['start_day'] = df['started_at'].dt.day\n","df['start_month'] = df['started_at'].dt.month\n","df['start_weekday'] = df['started_at'].dt.weekday"],"metadata":{"id":"fQ4C8zi3uvxk"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Add weather data\n","df = df.merge(weather_df, left_on=['start_day', 'start_month'], right_on=['start_day', 'start_month'], how='left')"],"metadata":{"id":"gjNBTE8jyzcG"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["### Calculate Edge Distances for GNN graph encoding"],"metadata":{"id":"ANHcjf9Bv8iZ"}},{"cell_type":"code","source":["# Haversine distance calculation derived from ChatGPT\n","\n","def haversine_distance(lat1: float, lon1: float, lat2: float, lon2: float) -> float:\n","    \"\"\"Calculate the great circle distance between two points on the earth.\"\"\"\n","    R = 6371  # Radius of the Earth in kilometers\n","\n","    lat1, lon1, lat2, lon2 = map(math.radians, [lat1, lon1, lat2, lon2])\n","    dlat = lat2 - lat1\n","    dlon = lon2 - lon1\n","\n","    a = math.sin(dlat/2)**2 + math.cos(lat1) * math.cos(lat2) * math.sin(dlon/2)**2\n","    c = 2 * math.asin(math.sqrt(a))\n","    return R * c"],"metadata":{"id":"G4nv9VCYsB9n"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Create Edges for GNN graph\n","edges = []\n","for region_name, (region_lat, region_lon) in REGION_LOCS.items():\n","    for other_region_name, (other_region_lat, other_region_lon) in REGION_LOCS.items():\n","        distance = haversine_distance(region_lat, region_lon, other_region_lat, other_region_lon)\n","        edges.append((region_name, other_region_name, distance))"],"metadata":{"id":"W8rw9iibtJ1k"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["edges_df = pd.DataFrame(edges, columns=['Region 1', 'Region 2', 'Distance'])\n","edges_df.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":206},"id":"c6DH9Cwvtmv7","executionInfo":{"status":"ok","timestamp":1731908637163,"user_tz":360,"elapsed":156,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"a163aa5c-430c-4c5e-cca7-e28ccb787773"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["   Region 1  Region 2  Distance\n","0         0         0  0.000000\n","1         0         1  5.034951\n","2         0         2  7.938928\n","3         0         3  8.643947\n","4         0         4  2.538970"],"text/html":["\n","  <div id=\"df-a17e4adc-13b3-497c-a611-06028d8d123b\" class=\"colab-df-container\">\n","    <div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>Region 1</th>\n","      <th>Region 2</th>\n","      <th>Distance</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0.000000</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>5.034951</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>0</td>\n","      <td>2</td>\n","      <td>7.938928</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>0</td>\n","      <td>3</td>\n","      <td>8.643947</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>0</td>\n","      <td>4</td>\n","      <td>2.538970</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>\n","    <div class=\"colab-df-buttons\">\n","\n","  <div class=\"colab-df-container\">\n","    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a17e4adc-13b3-497c-a611-06028d8d123b')\"\n","            title=\"Convert this dataframe to an interactive table.\"\n","            style=\"display:none;\">\n","\n","  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n","    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n","  </svg>\n","    </button>\n","\n","  <style>\n","    .colab-df-container {\n","      display:flex;\n","      gap: 12px;\n","    }\n","\n","    .colab-df-convert {\n","      background-color: #E8F0FE;\n","      border: none;\n","      border-radius: 50%;\n","      cursor: pointer;\n","      display: none;\n","      fill: #1967D2;\n","      height: 32px;\n","      padding: 0 0 0 0;\n","      width: 32px;\n","    }\n","\n","    .colab-df-convert:hover {\n","      background-color: #E2EBFA;\n","      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n","      fill: #174EA6;\n","    }\n","\n","    .colab-df-buttons div {\n","      margin-bottom: 4px;\n","    }\n","\n","    [theme=dark] .colab-df-convert {\n","      background-color: #3B4455;\n","      fill: #D2E3FC;\n","    }\n","\n","    [theme=dark] .colab-df-convert:hover {\n","      background-color: #434B5C;\n","      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n","      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n","      fill: #FFFFFF;\n","    }\n","  </style>\n","\n","    <script>\n","      const buttonEl =\n","        document.querySelector('#df-a17e4adc-13b3-497c-a611-06028d8d123b button.colab-df-convert');\n","      buttonEl.style.display =\n","        google.colab.kernel.accessAllowed ? 'block' : 'none';\n","\n","      async function convertToInteractive(key) {\n","        const element = document.querySelector('#df-a17e4adc-13b3-497c-a611-06028d8d123b');\n","        const dataTable =\n","          await google.colab.kernel.invokeFunction('convertToInteractive',\n","                                                    [key], {});\n","        if (!dataTable) return;\n","\n","        const docLinkHtml = 'Like what you see? Visit the ' +\n","          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n","          + ' to learn more about interactive tables.';\n","        element.innerHTML = '';\n","        dataTable['output_type'] = 'display_data';\n","        await google.colab.output.renderOutput(dataTable, element);\n","        const docLink = document.createElement('div');\n","        docLink.innerHTML = docLinkHtml;\n","        element.appendChild(docLink);\n","      }\n","    </script>\n","  </div>\n","\n","\n","<div id=\"df-0fd7372f-56dd-4d3e-8849-78025087b6d6\">\n","  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-0fd7372f-56dd-4d3e-8849-78025087b6d6')\"\n","            title=\"Suggest charts\"\n","            style=\"display:none;\">\n","\n","<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n","     width=\"24px\">\n","    <g>\n","        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n","    </g>\n","</svg>\n","  </button>\n","\n","<style>\n","  .colab-df-quickchart {\n","      --bg-color: #E8F0FE;\n","      --fill-color: #1967D2;\n","      --hover-bg-color: #E2EBFA;\n","      --hover-fill-color: #174EA6;\n","      --disabled-fill-color: #AAA;\n","      --disabled-bg-color: #DDD;\n","  }\n","\n","  [theme=dark] .colab-df-quickchart {\n","      --bg-color: #3B4455;\n","      --fill-color: #D2E3FC;\n","      --hover-bg-color: #434B5C;\n","      --hover-fill-color: #FFFFFF;\n","      --disabled-bg-color: #3B4455;\n","      --disabled-fill-color: #666;\n","  }\n","\n","  .colab-df-quickchart {\n","    background-color: var(--bg-color);\n","    border: none;\n","    border-radius: 50%;\n","    cursor: pointer;\n","    display: none;\n","    fill: var(--fill-color);\n","    height: 32px;\n","    padding: 0;\n","    width: 32px;\n","  }\n","\n","  .colab-df-quickchart:hover {\n","    background-color: var(--hover-bg-color);\n","    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n","    fill: var(--button-hover-fill-color);\n","  }\n","\n","  .colab-df-quickchart-complete:disabled,\n","  .colab-df-quickchart-complete:disabled:hover {\n","    background-color: var(--disabled-bg-color);\n","    fill: var(--disabled-fill-color);\n","    box-shadow: none;\n","  }\n","\n","  .colab-df-spinner {\n","    border: 2px solid var(--fill-color);\n","    border-color: transparent;\n","    border-bottom-color: var(--fill-color);\n","    animation:\n","      spin 1s steps(1) infinite;\n","  }\n","\n","  @keyframes spin {\n","    0% {\n","      border-color: transparent;\n","      border-bottom-color: var(--fill-color);\n","      border-left-color: var(--fill-color);\n","    }\n","    20% {\n","      border-color: transparent;\n","      border-left-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","    }\n","    30% {\n","      border-color: transparent;\n","      border-left-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","      border-right-color: var(--fill-color);\n","    }\n","    40% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","    }\n","    60% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","    }\n","    80% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","      border-bottom-color: var(--fill-color);\n","    }\n","    90% {\n","      border-color: transparent;\n","      border-bottom-color: var(--fill-color);\n","    }\n","  }\n","</style>\n","\n","  <script>\n","    async function quickchart(key) {\n","      const quickchartButtonEl =\n","        document.querySelector('#' + key + ' button');\n","      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n","      quickchartButtonEl.classList.add('colab-df-spinner');\n","      try {\n","        const charts = await google.colab.kernel.invokeFunction(\n","            'suggestCharts', [key], {});\n","      } catch (error) {\n","        console.error('Error during call to suggestCharts:', error);\n","      }\n","      quickchartButtonEl.classList.remove('colab-df-spinner');\n","      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n","    }\n","    (() => {\n","      let quickchartButtonEl =\n","        document.querySelector('#df-0fd7372f-56dd-4d3e-8849-78025087b6d6 button');\n","      quickchartButtonEl.style.display =\n","        google.colab.kernel.accessAllowed ? 'block' : 'none';\n","    })();\n","  </script>\n","</div>\n","\n","    </div>\n","  </div>\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"edges_df","summary":"{\n  \"name\": \"edges_df\",\n  \"rows\": 100,\n  \"fields\": [\n    {\n      \"column\": \"Region 1\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 2,\n        \"min\": 0,\n        \"max\": 9,\n        \"num_unique_values\": 10,\n        \"samples\": [\n          8,\n          1,\n          5\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Region 2\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 2,\n        \"min\": 0,\n        \"max\": 9,\n        \"num_unique_values\": 10,\n        \"samples\": [\n          8,\n          1,\n          5\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Distance\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 3.3372587528817186,\n        \"min\": 0.0,\n        \"max\": 13.473711301783453,\n        \"num_unique_values\": 46,\n        \"samples\": [\n          4.997551069111904,\n          6.759940433386638,\n          6.324017737363252\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"}},"metadata":{},"execution_count":51}]},{"cell_type":"markdown","source":["### Isolate Features and Labels"],"metadata":{"id":"KwvHbn0hwlnt"}},{"cell_type":"code","source":["df.drop(columns=['ride_id', 'started_at', 'ended_at', 'start_station_name', 'start_station_id', 'end_station_name', 'end_station_id'], inplace=True)\n","\n","# We exclusively use regions instead of lat/lng for more straightforward relationships between features and label\n","df.drop(columns=['start_lat', 'start_lng', 'end_lat', 'end_lng'], inplace=True)\n","df.dropna(inplace=True)\n","df.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":226},"id":"R8idMkgouxCg","executionInfo":{"status":"ok","timestamp":1731908641999,"user_tz":360,"elapsed":2109,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"4d8da9fd-f54c-4594-d73e-d014528b3df9"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["   rideable_type  member_casual  start_region  end_region  start_hour  \\\n","0              0              0             9           9          13   \n","1              1              0             5           5          18   \n","2              0              1             2           1           7   \n","3              0              0             6           6          10   \n","4              0              0             5           5          11   \n","\n","   start_day  start_month  start_weekday  PRCP  SNOW  SNWD  WESD  WESF  \n","0         15            2              2  0.00   0.0   0.0   0.0   0.0  \n","1         18           10              2  0.00   0.0   0.0   0.0   0.0  \n","2          9            3              3  0.00   0.0   0.0   0.0   0.0  \n","3          7            7              4  0.00   0.0   0.0   0.0   0.0  \n","4         10            9              6  0.35   0.0   0.0   0.0   0.0  "],"text/html":["\n","  <div id=\"df-478a73e3-f54a-4cbc-95ab-d5bf729f24e0\" class=\"colab-df-container\">\n","    <div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>rideable_type</th>\n","      <th>member_casual</th>\n","      <th>start_region</th>\n","      <th>end_region</th>\n","      <th>start_hour</th>\n","      <th>start_day</th>\n","      <th>start_month</th>\n","      <th>start_weekday</th>\n","      <th>PRCP</th>\n","      <th>SNOW</th>\n","      <th>SNWD</th>\n","      <th>WESD</th>\n","      <th>WESF</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>9</td>\n","      <td>9</td>\n","      <td>13</td>\n","      <td>15</td>\n","      <td>2</td>\n","      <td>2</td>\n","      <td>0.00</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>1</td>\n","      <td>0</td>\n","      <td>5</td>\n","      <td>5</td>\n","      <td>18</td>\n","      <td>18</td>\n","      <td>10</td>\n","      <td>2</td>\n","      <td>0.00</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>2</td>\n","      <td>1</td>\n","      <td>7</td>\n","      <td>9</td>\n","      <td>3</td>\n","      <td>3</td>\n","      <td>0.00</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>6</td>\n","      <td>6</td>\n","      <td>10</td>\n","      <td>7</td>\n","      <td>7</td>\n","      <td>4</td>\n","      <td>0.00</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>5</td>\n","      <td>5</td>\n","      <td>11</td>\n","      <td>10</td>\n","      <td>9</td>\n","      <td>6</td>\n","      <td>0.35</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>\n","    <div class=\"colab-df-buttons\">\n","\n","  <div class=\"colab-df-container\">\n","    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-478a73e3-f54a-4cbc-95ab-d5bf729f24e0')\"\n","            title=\"Convert this dataframe to an interactive table.\"\n","            style=\"display:none;\">\n","\n","  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n","    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n","  </svg>\n","    </button>\n","\n","  <style>\n","    .colab-df-container {\n","      display:flex;\n","      gap: 12px;\n","    }\n","\n","    .colab-df-convert {\n","      background-color: #E8F0FE;\n","      border: none;\n","      border-radius: 50%;\n","      cursor: pointer;\n","      display: none;\n","      fill: #1967D2;\n","      height: 32px;\n","      padding: 0 0 0 0;\n","      width: 32px;\n","    }\n","\n","    .colab-df-convert:hover {\n","      background-color: #E2EBFA;\n","      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n","      fill: #174EA6;\n","    }\n","\n","    .colab-df-buttons div {\n","      margin-bottom: 4px;\n","    }\n","\n","    [theme=dark] .colab-df-convert {\n","      background-color: #3B4455;\n","      fill: #D2E3FC;\n","    }\n","\n","    [theme=dark] .colab-df-convert:hover {\n","      background-color: #434B5C;\n","      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n","      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n","      fill: #FFFFFF;\n","    }\n","  </style>\n","\n","    <script>\n","      const buttonEl =\n","        document.querySelector('#df-478a73e3-f54a-4cbc-95ab-d5bf729f24e0 button.colab-df-convert');\n","      buttonEl.style.display =\n","        google.colab.kernel.accessAllowed ? 'block' : 'none';\n","\n","      async function convertToInteractive(key) {\n","        const element = document.querySelector('#df-478a73e3-f54a-4cbc-95ab-d5bf729f24e0');\n","        const dataTable =\n","          await google.colab.kernel.invokeFunction('convertToInteractive',\n","                                                    [key], {});\n","        if (!dataTable) return;\n","\n","        const docLinkHtml = 'Like what you see? Visit the ' +\n","          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n","          + ' to learn more about interactive tables.';\n","        element.innerHTML = '';\n","        dataTable['output_type'] = 'display_data';\n","        await google.colab.output.renderOutput(dataTable, element);\n","        const docLink = document.createElement('div');\n","        docLink.innerHTML = docLinkHtml;\n","        element.appendChild(docLink);\n","      }\n","    </script>\n","  </div>\n","\n","\n","<div id=\"df-7d146376-7e72-4837-a812-e080e7235f37\">\n","  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-7d146376-7e72-4837-a812-e080e7235f37')\"\n","            title=\"Suggest charts\"\n","            style=\"display:none;\">\n","\n","<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n","     width=\"24px\">\n","    <g>\n","        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n","    </g>\n","</svg>\n","  </button>\n","\n","<style>\n","  .colab-df-quickchart {\n","      --bg-color: #E8F0FE;\n","      --fill-color: #1967D2;\n","      --hover-bg-color: #E2EBFA;\n","      --hover-fill-color: #174EA6;\n","      --disabled-fill-color: #AAA;\n","      --disabled-bg-color: #DDD;\n","  }\n","\n","  [theme=dark] .colab-df-quickchart {\n","      --bg-color: #3B4455;\n","      --fill-color: #D2E3FC;\n","      --hover-bg-color: #434B5C;\n","      --hover-fill-color: #FFFFFF;\n","      --disabled-bg-color: #3B4455;\n","      --disabled-fill-color: #666;\n","  }\n","\n","  .colab-df-quickchart {\n","    background-color: var(--bg-color);\n","    border: none;\n","    border-radius: 50%;\n","    cursor: pointer;\n","    display: none;\n","    fill: var(--fill-color);\n","    height: 32px;\n","    padding: 0;\n","    width: 32px;\n","  }\n","\n","  .colab-df-quickchart:hover {\n","    background-color: var(--hover-bg-color);\n","    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n","    fill: var(--button-hover-fill-color);\n","  }\n","\n","  .colab-df-quickchart-complete:disabled,\n","  .colab-df-quickchart-complete:disabled:hover {\n","    background-color: var(--disabled-bg-color);\n","    fill: var(--disabled-fill-color);\n","    box-shadow: none;\n","  }\n","\n","  .colab-df-spinner {\n","    border: 2px solid var(--fill-color);\n","    border-color: transparent;\n","    border-bottom-color: var(--fill-color);\n","    animation:\n","      spin 1s steps(1) infinite;\n","  }\n","\n","  @keyframes spin {\n","    0% {\n","      border-color: transparent;\n","      border-bottom-color: var(--fill-color);\n","      border-left-color: var(--fill-color);\n","    }\n","    20% {\n","      border-color: transparent;\n","      border-left-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","    }\n","    30% {\n","      border-color: transparent;\n","      border-left-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","      border-right-color: var(--fill-color);\n","    }\n","    40% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","    }\n","    60% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","    }\n","    80% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","      border-bottom-color: var(--fill-color);\n","    }\n","    90% {\n","      border-color: transparent;\n","      border-bottom-color: var(--fill-color);\n","    }\n","  }\n","</style>\n","\n","  <script>\n","    async function quickchart(key) {\n","      const quickchartButtonEl =\n","        document.querySelector('#' + key + ' button');\n","      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n","      quickchartButtonEl.classList.add('colab-df-spinner');\n","      try {\n","        const charts = await google.colab.kernel.invokeFunction(\n","            'suggestCharts', [key], {});\n","      } catch (error) {\n","        console.error('Error during call to suggestCharts:', error);\n","      }\n","      quickchartButtonEl.classList.remove('colab-df-spinner');\n","      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n","    }\n","    (() => {\n","      let quickchartButtonEl =\n","        document.querySelector('#df-7d146376-7e72-4837-a812-e080e7235f37 button');\n","      quickchartButtonEl.style.display =\n","        google.colab.kernel.accessAllowed ? 'block' : 'none';\n","    })();\n","  </script>\n","</div>\n","\n","    </div>\n","  </div>\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"df"}},"metadata":{},"execution_count":52}]},{"cell_type":"code","source":["# Verify the new encoding\n","num_classes = len(df['end_region'].unique())\n","print(f\"Number of classes after encoding: {num_classes}\")\n","print(f\"Maximum class index after encoding: {df['end_region'].max()}\")"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"S9xM--Ahuycs","executionInfo":{"status":"ok","timestamp":1731908646147,"user_tz":360,"elapsed":143,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"2a25435b-e0e0-4ea0-fcc3-24f61ed16eb1"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Number of classes after encoding: 10\n","Maximum class index after encoding: 9\n"]}]},{"cell_type":"code","source":["# Separate features and target\n","features = df.drop(columns=['end_region'])\n","labels = df['end_region']\n","del df # Deleting unnecessary objects to save RAM"],"metadata":{"id":"pgB8jDMku6gd"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["features.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":226},"id":"SMYdpc0yu_dC","executionInfo":{"status":"ok","timestamp":1731908649456,"user_tz":360,"elapsed":168,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"8eaf945d-a27a-4e23-f4cd-4adf24cbb559"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["   rideable_type  member_casual  start_region  start_hour  start_day  \\\n","0              0              0             9          13         15   \n","1              1              0             5          18         18   \n","2              0              1             2           7          9   \n","3              0              0             6          10          7   \n","4              0              0             5          11         10   \n","\n","   start_month  start_weekday  PRCP  SNOW  SNWD  WESD  WESF  \n","0            2              2  0.00   0.0   0.0   0.0   0.0  \n","1           10              2  0.00   0.0   0.0   0.0   0.0  \n","2            3              3  0.00   0.0   0.0   0.0   0.0  \n","3            7              4  0.00   0.0   0.0   0.0   0.0  \n","4            9              6  0.35   0.0   0.0   0.0   0.0  "],"text/html":["\n","  <div id=\"df-b2c6b218-fa7a-41f5-b502-9cf0e28227da\" class=\"colab-df-container\">\n","    <div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>rideable_type</th>\n","      <th>member_casual</th>\n","      <th>start_region</th>\n","      <th>start_hour</th>\n","      <th>start_day</th>\n","      <th>start_month</th>\n","      <th>start_weekday</th>\n","      <th>PRCP</th>\n","      <th>SNOW</th>\n","      <th>SNWD</th>\n","      <th>WESD</th>\n","      <th>WESF</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>9</td>\n","      <td>13</td>\n","      <td>15</td>\n","      <td>2</td>\n","      <td>2</td>\n","      <td>0.00</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>1</td>\n","      <td>0</td>\n","      <td>5</td>\n","      <td>18</td>\n","      <td>18</td>\n","      <td>10</td>\n","      <td>2</td>\n","      <td>0.00</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>0</td>\n","      <td>1</td>\n","      <td>2</td>\n","      <td>7</td>\n","      <td>9</td>\n","      <td>3</td>\n","      <td>3</td>\n","      <td>0.00</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>6</td>\n","      <td>10</td>\n","      <td>7</td>\n","      <td>7</td>\n","      <td>4</td>\n","      <td>0.00</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>5</td>\n","      <td>11</td>\n","      <td>10</td>\n","      <td>9</td>\n","      <td>6</td>\n","      <td>0.35</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","      <td>0.0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>\n","    <div class=\"colab-df-buttons\">\n","\n","  <div class=\"colab-df-container\">\n","    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-b2c6b218-fa7a-41f5-b502-9cf0e28227da')\"\n","            title=\"Convert this dataframe to an interactive table.\"\n","            style=\"display:none;\">\n","\n","  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n","    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n","  </svg>\n","    </button>\n","\n","  <style>\n","    .colab-df-container {\n","      display:flex;\n","      gap: 12px;\n","    }\n","\n","    .colab-df-convert {\n","      background-color: #E8F0FE;\n","      border: none;\n","      border-radius: 50%;\n","      cursor: pointer;\n","      display: none;\n","      fill: #1967D2;\n","      height: 32px;\n","      padding: 0 0 0 0;\n","      width: 32px;\n","    }\n","\n","    .colab-df-convert:hover {\n","      background-color: #E2EBFA;\n","      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n","      fill: #174EA6;\n","    }\n","\n","    .colab-df-buttons div {\n","      margin-bottom: 4px;\n","    }\n","\n","    [theme=dark] .colab-df-convert {\n","      background-color: #3B4455;\n","      fill: #D2E3FC;\n","    }\n","\n","    [theme=dark] .colab-df-convert:hover {\n","      background-color: #434B5C;\n","      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n","      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n","      fill: #FFFFFF;\n","    }\n","  </style>\n","\n","    <script>\n","      const buttonEl =\n","        document.querySelector('#df-b2c6b218-fa7a-41f5-b502-9cf0e28227da button.colab-df-convert');\n","      buttonEl.style.display =\n","        google.colab.kernel.accessAllowed ? 'block' : 'none';\n","\n","      async function convertToInteractive(key) {\n","        const element = document.querySelector('#df-b2c6b218-fa7a-41f5-b502-9cf0e28227da');\n","        const dataTable =\n","          await google.colab.kernel.invokeFunction('convertToInteractive',\n","                                                    [key], {});\n","        if (!dataTable) return;\n","\n","        const docLinkHtml = 'Like what you see? Visit the ' +\n","          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n","          + ' to learn more about interactive tables.';\n","        element.innerHTML = '';\n","        dataTable['output_type'] = 'display_data';\n","        await google.colab.output.renderOutput(dataTable, element);\n","        const docLink = document.createElement('div');\n","        docLink.innerHTML = docLinkHtml;\n","        element.appendChild(docLink);\n","      }\n","    </script>\n","  </div>\n","\n","\n","<div id=\"df-363f27b6-dfa1-4010-912f-0251aafbaf35\">\n","  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-363f27b6-dfa1-4010-912f-0251aafbaf35')\"\n","            title=\"Suggest charts\"\n","            style=\"display:none;\">\n","\n","<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n","     width=\"24px\">\n","    <g>\n","        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n","    </g>\n","</svg>\n","  </button>\n","\n","<style>\n","  .colab-df-quickchart {\n","      --bg-color: #E8F0FE;\n","      --fill-color: #1967D2;\n","      --hover-bg-color: #E2EBFA;\n","      --hover-fill-color: #174EA6;\n","      --disabled-fill-color: #AAA;\n","      --disabled-bg-color: #DDD;\n","  }\n","\n","  [theme=dark] .colab-df-quickchart {\n","      --bg-color: #3B4455;\n","      --fill-color: #D2E3FC;\n","      --hover-bg-color: #434B5C;\n","      --hover-fill-color: #FFFFFF;\n","      --disabled-bg-color: #3B4455;\n","      --disabled-fill-color: #666;\n","  }\n","\n","  .colab-df-quickchart {\n","    background-color: var(--bg-color);\n","    border: none;\n","    border-radius: 50%;\n","    cursor: pointer;\n","    display: none;\n","    fill: var(--fill-color);\n","    height: 32px;\n","    padding: 0;\n","    width: 32px;\n","  }\n","\n","  .colab-df-quickchart:hover {\n","    background-color: var(--hover-bg-color);\n","    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n","    fill: var(--button-hover-fill-color);\n","  }\n","\n","  .colab-df-quickchart-complete:disabled,\n","  .colab-df-quickchart-complete:disabled:hover {\n","    background-color: var(--disabled-bg-color);\n","    fill: var(--disabled-fill-color);\n","    box-shadow: none;\n","  }\n","\n","  .colab-df-spinner {\n","    border: 2px solid var(--fill-color);\n","    border-color: transparent;\n","    border-bottom-color: var(--fill-color);\n","    animation:\n","      spin 1s steps(1) infinite;\n","  }\n","\n","  @keyframes spin {\n","    0% {\n","      border-color: transparent;\n","      border-bottom-color: var(--fill-color);\n","      border-left-color: var(--fill-color);\n","    }\n","    20% {\n","      border-color: transparent;\n","      border-left-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","    }\n","    30% {\n","      border-color: transparent;\n","      border-left-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","      border-right-color: var(--fill-color);\n","    }\n","    40% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","      border-top-color: var(--fill-color);\n","    }\n","    60% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","    }\n","    80% {\n","      border-color: transparent;\n","      border-right-color: var(--fill-color);\n","      border-bottom-color: var(--fill-color);\n","    }\n","    90% {\n","      border-color: transparent;\n","      border-bottom-color: var(--fill-color);\n","    }\n","  }\n","</style>\n","\n","  <script>\n","    async function quickchart(key) {\n","      const quickchartButtonEl =\n","        document.querySelector('#' + key + ' button');\n","      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n","      quickchartButtonEl.classList.add('colab-df-spinner');\n","      try {\n","        const charts = await google.colab.kernel.invokeFunction(\n","            'suggestCharts', [key], {});\n","      } catch (error) {\n","        console.error('Error during call to suggestCharts:', error);\n","      }\n","      quickchartButtonEl.classList.remove('colab-df-spinner');\n","      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n","    }\n","    (() => {\n","      let quickchartButtonEl =\n","        document.querySelector('#df-363f27b6-dfa1-4010-912f-0251aafbaf35 button');\n","      quickchartButtonEl.style.display =\n","        google.colab.kernel.accessAllowed ? 'block' : 'none';\n","    })();\n","  </script>\n","</div>\n","\n","    </div>\n","  </div>\n"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"dataframe","variable_name":"features"}},"metadata":{},"execution_count":55}]},{"cell_type":"code","source":["labels.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":241},"id":"G2C83TBT9tn9","executionInfo":{"status":"ok","timestamp":1731908649706,"user_tz":360,"elapsed":3,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"77a0e2e0-1428-424d-bdf5-b82c9126e81a"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0    9\n","1    5\n","2    1\n","3    6\n","4    5\n","Name: end_region, dtype: int64"],"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>end_region</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>9</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>5</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>6</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>5</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div><br><label><b>dtype:</b> int64</label>"]},"metadata":{},"execution_count":56}]},{"cell_type":"code","source":["# Split training and testing set\n","X_train, X_test, y_train, y_test = train_test_split(features, labels, test_size=0.2, random_state=42)\n","\n","# Convert to PyTorch tensors\n","X_train = torch.tensor(X_train.values, dtype=torch.float32, device=DEVICE)\n","X_test = torch.tensor(X_test.values, dtype=torch.float32, device=DEVICE)\n","y_train = torch.tensor(y_train.values, dtype=torch.long, device=DEVICE)\n","y_test = torch.tensor(y_test.values, dtype=torch.long, device=DEVICE)"],"metadata":{"id":"3pwgMMf8xBCq"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["## GNN + NN Fusion Model Code"],"metadata":{"id":"DI_IHhNoZ5hG"}},{"cell_type":"code","source":["# Edges for Graph\n","edge_index = torch.tensor(edges_df[['Region 1', 'Region 2']].values, dtype=torch.long, device=DEVICE).t().contiguous()\n","edge_weights = torch.tensor(edges_df['Distance'].values, dtype=torch.float32, device=DEVICE)\n","edge_weights.shape"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"sk6pOL_Z-rKL","executionInfo":{"status":"ok","timestamp":1731908655789,"user_tz":360,"elapsed":113,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"bd9d5478-c072-41f3-c6c7-07cec018d2aa"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["torch.Size([100])"]},"metadata":{},"execution_count":58}]},{"cell_type":"code","source":["# Nodes for graph\n","node_features = []\n","for key in range(10):\n","  (r1, r2) = REGION_LOCS[key]\n","  node_features.append([key, r1, r2]) # (label, latitude, longitude)\n","\n","node_features = torch.tensor(node_features, dtype=torch.float32, device=DEVICE)\n","node_features.shape"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"XVEhHy7uF6rE","executionInfo":{"status":"ok","timestamp":1731908656542,"user_tz":360,"elapsed":118,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"dd02befc-bf39-4826-d2be-94f246bb192c"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["torch.Size([10, 3])"]},"metadata":{},"execution_count":59}]},{"cell_type":"code","source":["# Create a list of Data objects for each batch of trips\n","batch_size = 8192\n","batched_data = []\n","\n","# Chunk trips and labels into batches\n","num_batches = len(X_train) // batch_size + (1 if len(X_train) % batch_size > 0 else 0)\n","\n","for i in range(num_batches):\n","    start_idx = i * batch_size\n","    end_idx = min((i + 1) * batch_size, len(X_train))\n","\n","    # Create a Data object for this batch\n","    batch_trips = X_train[start_idx:end_idx]\n","    batch_labels = y_train[start_idx:end_idx]\n","    batch_data = Data(nodes=node_features,\n","                      trips=batch_trips,\n","                      edge_index=edge_index,\n","                      edge_attr=edge_weights,\n","                      labels=batch_labels)\n","    batched_data.append(batch_data)\n","\n","# Create a DataLoader to manage the batched data\n","dataloader = DataLoader(batched_data, batch_size=1, shuffle=True)"],"metadata":{"id":"e98A8XItSitn"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["class HybridGCNModel(torch.nn.Module):\n","    def __init__(self, num_node_features, num_trip_features, num_nodes):\n","        super(HybridGCNModel, self).__init__()\n","        # GCN Layers\n","        self.conv1 = GCNConv(num_node_features, 16)\n","        self.conv2 = GCNConv(16, 16)\n","\n","        # MLP Layers for the trip features\n","        self.fc1 = torch.nn.Linear(num_trip_features, 128)\n","        self.fc2 = nn.Linear(128, 32) # Align dimensions with GCN output\n","\n","        # Final classifier layers\n","        self.fc3 = torch.nn.Linear(48, 32)\n","        self.fc4 = torch.nn.Linear(32, num_nodes)\n","\n","    def forward(self, data):\n","        x, edge_index, edge_weights, t = data.nodes, data.edge_index, data.edge_attr, data.trips\n","\n","        # GCN part: Graph-based learning\n","        gnn_out = F.relu(self.conv1(x, edge_index=edge_index, edge_weight=edge_weights))\n","        gnn_out = self.conv2(gnn_out, edge_index=edge_index, edge_weight=edge_weights)\n","\n","        # apply graph to trip based on start station\n","        start_station_embeddings = gnn_out[t[:, 2].long()]\n","\n","        # MLP part: Learning from trip features\n","        mlp_out = F.relu(self.fc1(t))\n","        mlp_out = self.fc2(mlp_out)  # Align with GCN output dimension\n","\n","        # Combine GCN and MLP outputs (concatenation)\n","        combined_out = torch.cat([start_station_embeddings, mlp_out], dim=1)\n","\n","        # Final classification\n","        out = F.relu(self.fc3(combined_out))\n","        out = self.fc4(out)\n","        output = F.log_softmax(out, dim=1)\n","\n","        return output"],"metadata":{"id":"Cp0sRIKMx4en"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["## Model Training"],"metadata":{"id":"J5x8kM1KVZe3"}},{"cell_type":"code","source":["# Model Initialization\n","num_node_features = 3  # Number of node features\n","num_trip_features = 12  # Number of trip features\n","num_nodes = 10  # Number of nodes in the graph\n","\n","model = HybridGCNModel(num_node_features, num_trip_features, num_nodes)\n","\n","# Loss and optimizer\n","criterion = nn.CrossEntropyLoss()\n","optimizer = optim.Adam(model.parameters(), lr=0.01)\n","model.to(DEVICE)\n","\n","# Training Loop\n","num_epochs = 50\n","for epoch in range(num_epochs):\n","    model.train()\n","    tot_loss = 0.0\n","    optimizer.zero_grad()\n","\n","    for i, batch in enumerate(dataloader):\n","      batch = batch.to(DEVICE)\n","\n","      out = model(batch)  # Forward pass (graph-based and feature-based)\n","      loss = criterion(out, batch.labels)  # Compute loss\n","      loss.backward()  # Backpropagate\n","\n","      torch.nn.utils.clip_grad_norm_(model.parameters(), max_norm=1.0)  # Gradient clipping\n","      optimizer.step()  # Update weights\n","      optimizer.zero_grad()\n","\n","      tot_loss += loss.item()\n","\n","    loss = tot_loss / len(dataloader)\n","    if epoch % 10 == 0: print(f'Epoch [{epoch+1}/{num_epochs}], Loss: {loss:.4f}')\n","\n","print(f'Final Loss: {loss:.4f}')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"outputId":"03cc885f-c6b6-4c93-fccf-dd227b153627","executionInfo":{"status":"ok","timestamp":1731908917448,"user_tz":360,"elapsed":257314,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"id":"0JYyV5KTuryP"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch [1/50], Loss: 1.4962\n","Epoch [11/50], Loss: 1.3730\n","Epoch [21/50], Loss: 1.3705\n","Epoch [31/50], Loss: 1.3696\n","Epoch [41/50], Loss: 1.3689\n","Final Loss: 1.3689\n"]}]},{"cell_type":"code","source":["# Evaluation (on test set)\n","model.eval()\n","test_data = Data(nodes=node_features,\n","            trips=X_test,\n","            edge_index=edge_index,\n","            edge_attr=edge_weights,\n","            labels=y_test)\n","with torch.no_grad():\n","    out = model(test_data)  # Forward pass on test set\n","    _, pred = torch.max(out, dim=1)  # Get the predicted classes\n","    accuracy = (pred == y_test).sum().item() / len(y_test)\n","    print(f'Test Accuracy: {accuracy:.4f}')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1731908926894,"user_tz":360,"elapsed":119,"user":{"displayName":"Aman Gupta","userId":"09473397613605643077"}},"outputId":"2714736f-456e-4228-fd66-f07fd808fa72","id":"j9QVNPj9uryQ"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Test Accuracy: 0.5548\n"]}]},{"cell_type":"markdown","source":["Our model yielded a final of **0.5548**. This is to be expected after examining our Data Exploration work - while there were slight relationships between the input features and end_regions, none were strong enough to suggest that one could reliably predict the end_region from start_region, weather, time, etc.\n","\n","Below is an explanation of all the variations to the model that we explored to try and improve the model:"],"metadata":{"id":"tUWM-HJvcITY"}},{"cell_type":"markdown","source":["# Exploring different training strategies"],"metadata":{"id":"XDSWP0AaaSSQ"}},{"cell_type":"markdown","source":["## Effects of mini-batch learning (Do not run)\n","\n","We tried the same model without mini-batch learning with the following code, but it ran out of RAM:"],"metadata":{"id":"BpZEfDG7aY3I"}},{"cell_type":"code","source":["# Create a PyTorch Geometric Data object\n","data = Data(nodes=node_features,\n","            trips=X_train,\n","            edge_index=edge_index,\n","            edge_attr=edge_weights,\n","            labels=y_train)"],"metadata":{"id":"Pb_0FueRelHR"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Model Initialization\n","num_node_features = 3  # Number of node features\n","num_trip_features = 7  # Number of trip features\n","num_nodes = 10  # Number of nodes in the graph\n","\n","model = HybridGCNModel(num_node_features, num_trip_features, num_nodes)\n","\n","# Loss and optimizer\n","criterion = nn.CrossEntropyLoss()\n","optimizer = optim.Adam(model.parameters(), lr=0.01)\n","model.to(DEVICE)\n","\n","# Training Loop\n","num_epochs = 300\n","for epoch in range(num_epochs):\n","    model.train()\n","    optimizer.zero_grad()\n","    out = model(data)  # Forward pass (graph-based and feature-based)\n","    loss = criterion(out, data.labels)  # Compute loss\n","    loss.backward()  # Backpropagate\n","\n","    torch.nn.utils.clip_grad_norm_(model.parameters(), max_norm=10.0)  # Gradient clipping\n","\n","    optimizer.step()  # Update weights\n","\n","    if epoch % 50 == 0:\n","      print(f'Epoch [{epoch+1}/{num_epochs}], Loss: {loss.item():.4f}')\n","\n","print(f'Final Loss: {loss.item():.4f}')"],"metadata":{"id":"EqmUHQzxxorx"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["With a smaller debugging dataset, the result was the model repeatedly converges to the similar loss values and final accuracy value. Thus we stuck to batch learning."],"metadata":{"id":"uatR5D0pdlGI"}},{"cell_type":"markdown","source":["We also tried different types of layers such as GATConv, Conv, and RNN. Each gave very similar losses/accuracies with worse memory results."],"metadata":{"id":"wh19_x_k0iy1"}},{"cell_type":"markdown","source":["## Effects of different optimizers\n","\n","We tried SGD and the Adam optimizer. Both optimizers yielded similar results."],"metadata":{"id":"GcASjLpAae5k"}},{"cell_type":"markdown","source":["## Effects of different hyperparameters"],"metadata":{"id":"Y0z4svK3aoS3"}},{"cell_type":"markdown","source":["We examined the result of using greater numbers of epochs to train with our debugging dataset. As can be seen in the graph below, for the initial 0-50 epochs, there was a sharp decrease in loss and stabilized at around 1.8-1.9 loss. Then from 50-150 epochs there were some minor oscillations in the loss but overall, a slow decrease. From 150-300, the loss remained very stable with slight fluctuations which might be a result of overfitting. Thus we trained with 300 epochs (there is no point in training further than that).\n","\n","![image.png](data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAioAAAHHCAIAAAAeYOQlAAAAAXNSR0IArs4c6QAAAERlWElmTU0AKgAAAAgAAYdpAAQAAAABAAAAGgAAAAAAA6ABAAMAAAABAAEAAKACAAQAAAABAAACKqADAAQAAAABAAABxwAAAADK1h5DAABAAElEQVR4Ae2dB3xVRd6/SU+ANAKhlyAC0hQpFqyrouhrL2tZ67vrqth21/2vrGt/XdTdtb8va9dV1FddUV8VEV0bq5AAFkAF1NCRGhJCSM//e+8kh8tNu0nuufecc58jn+ucmTkzv3nm5HzPlDMTV1dX14kDAhCAAAQgEFkC8ZHNjtwgAAEIQAACPgLID/cBBCAAAQhEgQDyEwXoZAkBCEAAAsgP9wAEIAABCESBAPITBehkCQEIQAACyA/3AAQgAAEIRIEA8hMF6GQJAQhAAALID/cABCAAAQhEgQDyEwXoZAkBCEAAAsgP90CUCVxyySWDBg0K0YjbbrstLi4uxMhEcxEBU7Nbt251kc2Y2kECyE8HAcbi5RKAFo6PPvrIpVAkhF27do2i8VoB67nnnjviiCOysrI6d+48evToO+64Y9euXZE0ychA4/r96aefImkGecUCgcRYKCRlDC8BPSKtBP/xj3/MnTs30Ge//fazQkNxPP7447W1taHEVJw//elPN954Y4iR3RWtpqbm/PPPf/nllw8//HBpgOTn008/vf3221955ZX333+/Z8+ekSzOjBkzgpRYihhJA8grFgggP7FQy2Eu4y9+8Qsrxfnz50t+An2sIDnKysr0GA30aexOSkpq7NmcT6L/aC7U1f733nuvtOeGG274y1/+Ygpy+eWXn3POOaeddpqaZbNnzw576VqonbPOOqt79+5hz5EEIRBIgM63QBq4O0rgqKOOGjVq1KJFi9SDJOH54x//qBTfeOONk046qU+fPikpKfvss8+dd96pN30rp8Cxn1WrVqnb569//etjjz2mmIo/YcKEgoICK3LQ2I8iX3311a+//royVeSRI0e+++67VmQ51BM4fvz41NRUpfboo48GXR4Ys1W3WiHjxo1LS0vTc1lyu379eusSdUxdeuml/fr1kw29e/c+9dRTVRATunDhwuOPP16X6MK8vLzLLrvMuirQsXv3bqnO0KFDp0+fHuh/8sknX3zxxSqUZF7+//Ef/zF48ODACHIfcsghKqPl+fzzzxs7u3Xrdu65565du9YKarJ2rNBQHOIp5v/7v/+rmu3Vq1eXLl1OOeWUwCyUSAugFPrdd99JU3v06CEgw4YNu+mmmwLz3bFjh+4HtbQyMzOFVAJpheot57DDDlOQmmW60NxaVigONxKg9ePGWnO0zdu2bZsyZYoefHpGmy6jZ555Ro+M3/72t/r917/+dcstt5SUlFjv+I0L88ILL+zcufPXv/61nnRqE5xxxhk//vhjc42kefPmvfbaa1dddVV6evpDDz105plnrlmzJicnR8l+8cUXJ5xwgvRAXVgSPI2j6KnXOLtQfFQEPQ2lhZKHTZs2Pfjgg//+97+VvumSUqbLli275pprNIdi8+bNelDKBuOePHmyMlWHoWJKk2Rqk9mpFEVFRdddd51ad0ERLrrooqeffvqtt946+OCDf/7zn+tUeixLTLTVq1dLmSyYd911180336zn+y9/+cstW7Y8/PDDeg+w7NQljWsnKDtzun379kB/WWVKajyVi6rmD3/4gwr7wAMPHHvssV9++aXkRKEtg/r666/VtaiqVMNOfH744Yf/+7//U2pWXrJcIi3IixcvfuKJJ3Jzc++55x6FCq+kd8yYMapEafz3338v/tZVONxKQKOdHBBoN4GpU6fq1rcuP/LII3X697//3fKRQ++wgafSFTWMysvLjafe7gcOHGjchYWFulziocef8VHLST56SJnTW2+9NTA7uZOTk/UwMqFfffWVfPTMNadqOigjNVPM6cqVK83D3Zw2/pUlep1v7F9ZWannoBpYaqOYUImBMpKO6lSyIbcEoPGFs2bNUpDUonFQkI8e4oqp+EH+OjVKIA2Wu7i4WA/f3/3ud1Y0ybOUQCIkH8lbQkKCnuZW6JIlS1Rky6fJ2rEiG4chLGMCD7U2TOiHH34o/759++oFwviow1A+0mOdtgxKEaSFeksw1prLNexnHCZftQ7NqX5PP/103Qnm9P7771cuElQrFIcHCND5FvhXhjsMBPR8VEMhMCHzXiwftWk0s1bvvxIkdcIExgl06x0/Ozvb+CiyHGr9BEYIdOvVWx1rxkdvxxkZGSaymjsasdfAiTr9TOiQIUPULAu8NkS3OtD0mq8GljrxzCXqSxw+fPjbb7+tU5VOEqheKaNDgWmaFoO0qqqqKtC/sVtk5KlHc+Mg46nHvYJUOhVBT3w9ekxM9YOpVTRgwACdqmmlp7kaEIJsDvWP7bvvvkYzTPzGtWP8g37/+c9/qg1nHWp+BUZQC8wyVaNEal++8847itAyKInHJ598IoEx1poEpZ2BKV9xxRXWqapebTVTcENS7yKhz1Kx0sHhWALIj2Orxq2G6dVYj+NA69VzojdZ9ebr6ameKDNPQS/ygXEC3YGPJ6NDjZ/sVvzAyPJUfBNZgqHGiiTHiilH0GlgUAtuva0rVC2AwDiSH+OvB7o6iDQ1QD2NertXc8Sao6zWhvrl1PWnsR8NCOkhXlFREZiI5TZPcyNClqdxBCmTtFljLZ9//rlC1XmlYTb5mJhq3kmWpDeCbB3ffvutUFhpNq4dKyjQoYJI161Dw0uBocrCOpV+iKoZ62oZlHktUCPSuraxI7A2A6teZZw0aZJ6FAVZ/boSYHSoMT3X+SA/rqsypxtstXWMoRpM1lNY3WLqtVcfml6oTW9+C48P9SAFFdJ62Q/y12mbIje+PCw+119//YoVKzRioeaRhl409VzDLUpZj+ZXX31VUqH5EeoD1Iu/JgWUlpY2ztTMVtfQSOMg4zlixAgTZHoUTZeXfuPj488++2wTJKTKUfMUrFaLcWjOhZVsUO1Y/g5xNFebMlstJzVnL7zwQgGRGh133HGBE1gcYj9mtIkA8tMmXERuMwH1SqkLRSPSGlfX6LFeqK2OtTan1ZYLNFojMdCwUOBFQaeBQS24NTSl0OXLlwfG0anxN57qANSQzHvvvbd06VINgfztb3+zIqtzTKMv6piaOXOmGoIvvfSSFWQ5zJwuzblo/EjVl1WKJnQmskan5NbsMomNet7UQ2X1LsoG6bSG7q1Wi3HIACujsDjUzLLSUY6iqnkE8mkZlJmzJz7WtW1ySGiPOeaY++6775tvvhFPzWEJ7FRsU1JEdggB5MchFeFZM8z7rNV80aP5f/7nfyJQWuWrh6/mZG/YsMFkp6dk+76e0bRmiZnmU1hdZ0pHnVoaAVLKGsfSNAqrRNIA9aSZmOoGtAquCAcccIB+rUSsS+TQFAl98SNJC5qIrOElKbembgdKiN79VShNDFOb0up5UyKanqBSq68vMFO5Jf+BeXXcLUW0+gnVvNu4caMZVGsZlPoD1af31FNPaVqgZUOgqZZnY0fQTLwWSDa+Fh/HEgie5elYQzHMpQQOPfRQNXc0qezaa69V15DWRwjxidPx8uorHzVHNGZw5ZVXqlXxyCOPaOBBU4RbSFlzBP7rv/4rMIK+ntGkA3UYaj6FehHPO+88M/Fa7/u/+c1vFFPdbnor14C/+sc0zUyz1xRB4xMKevbZZ6W1GveSJul5rfUdNPp14oknBqZvuTU5W112ykiddRoxUneTZmPrIx71yykdK5ocSkEKJ7mS2CimFaRcZPy0adM0EqM5F4qjmYSyR7OcFdmKFopDoqJZ8oEx1dllptHLU0zUXBMQlVRz9jT286tf/Ur+mlHdAihF0Mx4XXjggQfKJLXSZKf0teUaMTao51adb9J7NbA0lCWq+spKSZlQft1KQM8CDgi0m0Djidf69jMoNX2ioZd3PU/VTfT//t//mzNnjv5a1HNiojWeeB00iVmRNSvXRG488VoGmCDzq8eTErR8Pvjgg7Fjx2oqhB7Nai6of0w9clZokEMXNv4z1oUmmnq6lJQmGujhe8EFF6xbt874a46ZbNBMBHWLaXrFQQcdZGamKVQfr0iuNJyuq9R+UqeZuuCCMg08lUZqeoL0UiolO0VSTRmNFQXGMW4ZIFPVvGscpElrei7LGB2ySrapUWWiST4b105QCk1OvLbqy/R3vfjiixI5lUh1KkkInEit1JoDZTJS55v0WDPZVEDN5tBQmfE3+QZOrTbT7aSgiqB61NwN3T+qSv2KqlTfXMivewnEyfTGf3L4QMCTBNQm0OhL4NCFJ4tpX6E0knf00Udr5Enzre3LhZRjhABjPzFS0TFaTM29tkou1dHnKVp4xvLBAQEIRJEAYz9RhE/WthPQbCutIaZfdRBpFWd13aj3z/ZcyQACEAiBAPITAiSiuJaA1nzTQIW+A9Xoi76d/POf/xz4yaRri4XhEPACAcZ+vFCLlAECEICA6wgw9uO6KsNgCEAAAl4ggPx4oRYpAwQgAAHXEXDW2I/WEdHn3PpcTt8nug4lBkMAAhCAgEVAX/Xoa2t9p6UFkyzPQIez5Efa079//0D7cEMAAhCAgHsJaIF2LVHRpP3Okh+1e2SlzNVX302aiycEIAABCLiCgPZqUnPCPNWbNNhZ8mP63KQ9yE+TtYUnBCAAAXcRaGEkpekuOXcVD2shAAEIQMB1BJAf11UZBkMAAhDwAgHkxwu1SBkgAAEIuI4A8uO6KsNgCEAAAl4ggPx4oRYpAwQgAAHXEUB+XFdlGAwBCEDACwSQHy/UImWAAAQg4DoCyI/rqgyDIQABCHiBAPLjhVqkDBCAAARcRwD5cV2VYTAEIAABLxBAfrxQi5QBAhCAgOsIID+uqzIMhgAEIOAFAh6Un5+Ky3/cUuqFyqEMEIAABLxLwGvy88y/Cw+e/sFf31vu3SqjZBCAAAS8QMBr8jOiT6aqJb9wuzba80L9UAYIQAACHiXgNfkZ0y8zOTF+a2ll4dZdHq0yigUBCEDACwTslR9t9H399dcPHDgwLS3t0EMPLSgosJtZalLCAf2ylIsaQHbnRfoQgAAEINBuAvbKzy9/+cu5c+c+99xzS5YsmTx58rHHHrt+/fp22xrihRPzuilm/irkJ0RgRIMABCAQBQI2ys/u3bv/+c9/3nvvvUccccSQIUNuu+02/c6YMcPuUtbLD60fu0GTPgQgAIEOEEjswLWtXFpdXV1TU5OammrFUxfcvHnzrFPjqPAfxl1SUhIU2o7TAwdmx8d1Wle0e8OO3X2y0tqRApdAAAIQgIDdBGxs/aSnpx9yyCF33nnnhg0bpEPPP//8559/vnHjxqAiTZ8+PbPh6N+/f1BoO067piSO6uub/1ZA/1s78HEJBCAAgYgQsFF+ZL9GfTQBum/fvikpKQ899NB5550XHx+c47Rp04objrVr14al1BMG+YZ/FtD/FhaaJAIBCEDABgLBYhDeLPbZZ5+PP/64tLRUupKfn19VVTV48OCgLKRMGQFHUGj7Ts3wTwHy0z58XAUBCEDAfgL2yo+xv0uXLr179y4qKpozZ86pp55qf6E6mdbPys2l23dVRiA7soAABCAAgbYSsFd+pDfvvvtuYWGhpl8fffTRw4cPv/TSS9tqYjvid+uSvG9uV13I8E876HEJBCAAgQgQsFd+NKYzdepUqc5FF1102GGHSY2SkpIiUCplMcF8/UP/W2RwkwsEIACBNhKwceK1LDnHf7TRpPBEPyiv2wsL1rD2QXhokgoEIACBcBOwt/UTbmvbkJ4Z/lm2obi0oroNlxEVAhCAAAQiQsCz8qMPTvtlp9XWdVq0uigiJMkEAhCAAATaQMCz8iMGTL9uw41AVAhAAAKRJeBp+fF/fMrwT2TvKHKDAAQgEBIBT8uPf/Lbl+t2lFfVhASDSBCAAAQgECkCXpafvO5dundNrqyu/XpdcaR4kg8EIAABCIREwMvyExcX17D5wraQYBAJAhCAAAQiRcDL8iOGE83wzyomv0XqhiIfCEAAAqER8Lj8mLUPFq3aXl1TGxoQYkEAAhCAQCQIeFx+hvfKSE9N3FVZ8+3GnZHASR4QgAAEIBAaAY/LT0J83PiB2UKxoJDhn9DuCGJBAAIQiAgBj8uPGE7My9EvS19H5HYiEwhAAAKhEogF+fG1fgpWFWnf1VCpEA8CEIAABGwm4H35Gd03KyUxXvvO/bCl1GaYJA8BCEAAAqES8L78JCfGHzjADP9sD5UK8SAAAQhAwGYC3pcfAWTrOZvvIpKHAAQg0GYCMSE/2npOYLT2KMM/bb5BuAACEICAPQRiQn7GDshKjI/bWFy+rmi3PRhJFQIQgAAE2kYgJuSnc3LiqL6ZAsP067bdHcSGAAQgYBuBmJAf0WtYe5TZB7bdSiQMAQhAoC0EYkZ+6tceRX7acncQFwIQgIBtBGJFfiYM6hYX1+nHLbu27KywDSYJQwACEIBAqARiRX4yOycN65kuKgz/hHprEA8CEICAnQRiRX7EkOEfO28k0oYABCDQNgIxJD/qfxMbff3TNkLEhgAEIAABGwjEkPyY1s+3P5WUlFfZQJIkIQABCECgDQRiSH56ZqQOzOmsZa8Xsfd2G+4QokIAAhCwhUAMyY/4TWT6tS13EYlCAAIQaDOB2JIf1h5t8w3CBRCAAATsIRBb8mPWHv163Y7yqhp7eJIqBCAAAQiERCC25GdAt849M1Kqauq+WLMjJDxEggAEIAABewjElvzExcUx/dqeG4lUIQABCLSNQGzJj9iY/jfWPmjbbUJsCEAAAuEmYK/81NTU3HzzzXl5eWlpafvss8+dd94Z9Q3fzOyDRauLqmpqww2T9CAAAQhAIFQCiaFGbFe8e+65Z8aMGc8+++zIkSMXLlx46aWXZmZmXnvtte1KLDwXDc1Nz0xLKt5dtWxDyQH9s8KTKKlAAAIQgEAbCdjb+vnss89OPfXUk046adCgQWedddbkyZPz8/PbaGGYo8fHa/gnW4nmF24Lc9IkBwEIQAACIROwV34OPfTQDz74YMWKFbLnq6++mjdv3pQpU4Jsq6ioKAk4gkLtOG1Ye7TIjsRJEwIQgAAEQiFgb+fbjTfeKGUZPnx4QkKCxoHuuuuuCy64IMis6dOn33777UGetp5OzMtR+pp9UFtbp8aQrXmROAQgAAEINEnA3tbPyy+/PHPmzBdeeGHx4sUaAfrrX/+q3yA7pk2bVtxwrF27NijUjtORfTLSkhI0/LNi80470idNCEAAAhBolYC9rZ/f//73agCde+65smP06NGrV69WW+fiiy8ONCvFfwT62O1OSogfNzB73vdbCwq3D++VYXd2pA8BCEAAAo0J2Nv6KSsri4/fk4W64GprHTHd2Xx8uoC9fxrfEfhAAAIQiAgBe1s/J598ssZ7BgwYoInXX3zxxX333XfZZZdFpFytZGJmH2j4R98haSmEVmITDAEIQAAC4SZgr/w8/PDD+uz0qquu2rx5c58+fX7961/fcsst4S5Ce9IbOyArKSFuU0nFmu1lA3O6tCcJroEABCAAgQ4QsFd+0tPTH/AfHbDQlktTkxLG9MvS2gfaexv5sQUxiUIAAhBokcCegZkWo3kwsOHrn+0eLBtFggAEIOB4AjEsP+x86vi7EwMhAAEPE4hd+Rk3KFtzDlZvK9tUUu7hCqZoEIAABJxJIHblJyM1aT//Rz8a/nFm3WAVBCAAAQ8TiF35UaVa0689XMEUDQIQgIAzCSA/Wvqa1o8zb06sggAEvEwgpuXHrH2wfNPOHWWVXq5kygYBCEDAeQRiWn56pKcM7t6lrq7TwlVsvuC8exOLIAABTxOIaflRzdZ//bOK/jdP3+YUDgIQcB4B5KebKoXhH+fdmVgEAQh4nECsy48Z/lm6vrisstrjVU3xIAABCDiJQKzLT7/stD6ZqdW1dV+s2eGkesEWCEAAAh4nEOvyo90WJuT5+t/Y+8fjdzrFgwAEHEYg1uVH1VH/8Slf/zjs1sQcCEDA2wSQn04T/WuPLl5TVFntiJ1YvX3DUToIQAAChgDy02lIbtduXZIrqmuXrC/mtoAABCAAgcgQQH46+YZ/BmULN9OvI3PPkQsEIAABEUB+fLeBmX6dX7iNewICEIAABCJDAPnxcT4oL0e/C1cX1dTWRYY7uUAAAhCIcQLIj+8G2K93epfkhJ3l1d/9VBLjNwTFhwAEIBAZAsiPj3NiQvw4//y3AqZfR+a+IxcIQCDmCSA/9bfARDP7gLVHY/5PAgAQgEBkCCA/9Zwn+od/8guL6rQBAwcEIAABCNhMAPmpBzymX2ZyYvzW0orCrbtsZk7yEIAABCDAxOuGeyA1KeGAflk64+ufBiT8HwIQgICNBGj97IHL1nN7WOCCAAQgYDMB5GcPYLP0Na2fPURwQQACELCNAPKzB+24gdnxcZ3WFe3esGP3Hl9cEIAABCBgAwHkZw/UrimJI/tk6ryA6dd7qOCCAAQgYAsB5GcvrPXDP3x8uhcVTiAAAQiEnwDysxdT5GcvHJxAAAIQsI0A8rMXWrP09crNpdt3Ve4VwAkEIAABCISVgL3yM2jQIO2mE3hMnTo1rPaHOTHtO7dvblclyvBPmMmSHAQgAIG9CdgrPwUFBRsbjrlz5yrrs88+e28DHHfG9GvHVQkGQQACXiRgr/z06NGjV8Px1ltv7bPPPkceeaTDMR6U100W0vpxeDVhHgQg4HYCiZEpQGVl5fPPP//b3/5WHXFBOVb4D+NZUhL97XbM8M/S9cWlFdWaih1kLacQgAAEIBAWAva2fiwTX3/99R07dlxyySWWj+WYPn16ZsPRv39/yz9ajj5Zaf2y07Tr6eLVRdGygXwhAAEIeJ5AhOTnySefnDJlSp8+fRoDnTZtWnHDsXbt2sYRIu8z0b/1HKvvRJ48OUIAArFDIBKdS6tXr37//fdfe+21JrGm+I8mg6Llqa9/XvtiPfITLf7kCwEIxAKBSLR+nn766dzc3JNOOsktQM3Hp1+u21FeVeMWm7ETAhCAgLsI2C4/tbW1kp+LL744MTESLa2w0M/r3qV71+TK6tqv1xWHJUESgQAEIACBIAK2y4+63dasWXPZZZcFZezkU03PMw0gpl87uZqwDQIQcDUB2+Vn8uTJdXV1Q4cOdRcmM/16AWuPuqvasBYCEHAPAdvlxz0o9rLUtH4097q6pnavAE4gAAEIQCAcBJCfpikO75WRnpKoL0+/3biz6Rj4QgACEIBABwggP03DS4iPGz8oW2H5bD3XNCF8IQABCHSIAPLTLL6JeTkKyy/c1mwMAiAAAQhAoL0EkJ9myU3M87V+ClYVaepEs5EIgAAEIACBdhFAfprFNrpvVkpivPad+2FLabORCIAABCAAgXYRQH6axZacGD92QJaCmX7dLCMCIAABCLSXAPLTEjkz/FPA1z8tQSIMAhCAQHsIID8tUWPp65boEAYBCECgAwSQn5bgHTgwKzE+bkNx+bqispbiEQYBCEAAAm0kgPy0BKxzcuKovpmKweYLLWEiDAIQgEDbCSA/rTAzq+8gP61gIhgCEIBAGwkgP60Aqx/+Ye2DVjgRDAEIQKBtBJCfVniZpXd+3LJry86KVqISDAEIQAACIRNAflpBldU5eXivdEVaSAOoFVQEQwACEGgDAeSndVjs/dM6I2JAAAIQaCMB5Kd1YOx82jojYkAAAhBoIwHkp3VgRn6+2VhSUl7VemxiQAACEIBACASQn9Yh9cxIHZjTWcteL1pV1HpsYkAAAhCAQAgEkJ8QIHXqxPTrkDARCQIQgEDIBJCfkFBNyOumeHx8GhIsIkEAAhAIgQDyEwKkTp0O8svP1+t2lFfVhHQBkSAAAQhAoEUCyE+LeBoCB3TrnJueUlVT98WaHQ1+/B8CEIAABNpPAPkJiV1cXBzTr0MiRSQIQAACoRFAfkLjpNkHDP+Eiop4EIAABFongPy0zsjEMPKzeE1RVU1tqNcQDwIQgAAEmiGA/DQDppH30Nz0zLSkssqaZRtKGgXiAQEIQAACbSOA/ITKKz4+bsKgbMXOL9wW6jXEgwAEIACBZgggP82Aacq7YfiHtQ+aooMfBCAAgbYQQH7aQMssfV2wanttbV0bLiMqBCAAAQg0IoD8NELSvMeovhr9SSjeXbVyc2nzsQiBAAQgAIHWCSA/rTOyYiQlxB84MEunDP9YTHBAAAIQaB8B5Kdt3CYOytEF+Sx93TZsxIYABCAQTMB2+Vm/fv0vfvGLnJyctLS00aNHL1y4MNgEV503zD7YVqcNGDggAAEIQKC9BBLbe2FI1xUVFU2aNOnoo4+ePXt2jx49Vq5cmZ3tm7vs3mPsgKykhLhNJRVrtpcNzOni3oJgOQQgAIHoErBXfu65557+/fs//fTTppB5eXnRLW3Hc09NShjTL2vR6iJtvoD8dJwnKUAAAjFLwN7OtzfffHP8+PFnn312bm7u2LFjH3/88cagKyoqSgKOxhGc5mOmX7P3j9PqBXsgAAF3EbBXfn788ccZM2bsu+++c+bMufLKK6+99tpnn302CND06dMzGw41lYJCHXhq9v7R1z8OtA2TIAABCLiFQJytQ+jJyclq/Xz22WcGh+SnoKDg888/D6Sj1o8O46NWkBSouLg4IyMjMI6j3Pru54A73tPMg/w/HpObkeoo2zAGAhCAgEMI6HmulkULz3N7Wz+9e/ceMWKExWK//fZbs2aNdWocKSkpEhvrCAp14Kk+Pd2vl08d82kAObB6MAkCEHAJAXvlR9Peli9fbqFYsWLFwIEDrVP3OhqmX9P/5t46xHIIQCDKBOyVn9/85jfz58//85///P3337/wwguPPfbY1KlTo1zicGSP/ISDImlAAAIxTcBe+ZkwYcKsWbNefPHFUaNG3XnnnQ888MAFF1zgAd5m8tvyTTt3lFV6oDgUAQIQgEDkCdj73Y/K8x/+I/IFszXHHukpg7t3+XHrroWrio4d0dPWvEgcAhCAgCcJ2Nv68SQyUyjT/8b0aw9XMUWDAARsJYD8tBOv6X9bUMjsg3YC5DIIQCDGCSA/7bwBTOtn6frissrqdibBZRCAAARimADy087K75ed1jsztbq27os1O9qZBJdBAAIQiGECyE87Kz8uLs40gOh/aydBLoMABGKbAPLT/vqvn33A8E/7EXIlBCAQuwSQn/bX/cRB3XTx4jVFldW17U+FKyEAAQjEJAHkp/3VPiS3a7cuyRXVtUvWF7c/Fa6EAAQgEJMEkJ/2V7uGf8YP9G3eyt4/7YfIlRCAQKwSQH46VPN8fNohfFwMAQjEMAHkp0OVb8lPTW1dhxLiYghAAAIxRgD56VCFj+id0SU5YWd59fKfdnYoIS6GAAQgEGMEQpWftWvXrlu3zsDJz8+//vrrtXtCjLFqoriJCfHj/PPf8gu3NRGMFwQgAAEINEMgVPk5//zzP/zwQyXy008/HXfccVKgm2666Y477mgm2RjynjjIP/uAnU9jqM4pKgQgEAYCocrP0qVLJ06cqAxffvllbd7z2WefzZw585lnngmDCS5PYmJejkqQX1hUV8fwj8vrEvMhAIEIEghVfqqqqlJSUmTY+++/f8opp8gxfPjwjRs3RtBUh2Y1pl9mckL81tKKwq27HGoiZkEAAhBwHoFQ5WfkyJF///vfP/3007lz555wwgkqyIYNG3JyfC/+MX6kJiUc0D9LENj7J8bvBIoPAQi0iUCo8nPPPfc8+uijRx111Hnnnbf//vsrjzfffNN0x7UpP09GnpDnG/5h7VFPVi6FggAEbCIQ6mbbEp6tW7eWlJRkZ/setTouv/zyzp07G3eM/2r4578//IHWT4zfBhQfAhBoE4FQWz+7d++uqKgw2rN69eoHHnhg+fLlubm5bcrMq5HHDcyOj+u0dvvuDTt2e7WMlAsCEIBAeAmEKj+nnnrqP/7xD+W9Y8eOgw466G9/+9tpp502Y8aM8Frj0tS6piSO7JMp42kAubQGMRsCEIg8gVDlZ/HixYcffrjse/XVV3v27KkGkNTooYceirzFzszRrL7D2qPOrB2sggAEHEggVPkpKytLT09XAd57770zzjgjPj7+4IMPlgg5sEhRMWlC/doH26OSO5lCAAIQcB2BUOVnyJAhr7/+upbemTNnzuTJk1XOzZs3Z2RkuK7ANhk8wb/2wcrNpdt3VdqUBclCAAIQ8BKBUOXnlltuueGGGwYNGqTJ1occcogQqBk0duxYL7HoSFlyuqZo9zmlwPBPRzByLQQgEDsEQpWfs846a82aNQsXLlTrx9A55phj7r///tgh1WpJ6zdfKKT/rVVURIAABCDQKdTvfoSql/8w617369ePb06Dbp+D8rq9sGBNPmuPBnHhFAIQgEBTBEJt/dTW1mp968zMzIH+Iysr684775RnU2nGqJ+ZfbB0fXFpRXWMIqDYEIAABEImEGrrR9srPPnkk3ffffekSZOU+Lx582677bby8vK77ror5Lw8HrFPVlq/7LR1RbsXry46YmgPj5eW4kEAAhDoGIFQ5efZZ5994oknzFrXynHMmDF9+/a96qqrkJ9A/hMHdVtXtF5f/yA/gVhwQwACEGhMINTOt+3bt2uHhcDrdSrPQB/c9R+fMvzDrQABCECgNQKhyo9WuX7kkUcCU9Op2kCBPrgn5HUThC/X7qioroEGBCAAAQi0QCDUzrd77733pJNO0l5z5qOfzz//XJ+gvvPOOy0krSCND91+++1WnGHDhn333XfWqfccg7t36d41eWtp5dfris1MBO+VkRJBAAIQCAuBUFs/Rx555IoVK04//XQtOapD6+4sW7bsueeea9UI7VOnTVHNoQkLrcZ3dYS4uDhW33F1DWI8BCAQMQKhtn5kUJ8+fQInGnz11VeaC/fYY4+1bGtiYqK+F2o5jpdCNfwze+lPmn0w9WgvFYuyQAACEAgzgVBbP+3OduXKldKtwYMHX3DBBVo3oXE62kZIu9hZR+MI7vIxsw8WrS6qruGjKHdVHdZCAAIRJWCv/GhnoGeeeebdd9/VzkCFhYXasmHnzp1B5Zs+fbq+ZjVH//79g0Jddzq8V0Z6SqK+PP12Y3BJXVcWDIYABCBgHwF75WfKlClnn322Jsgdf/zxmqegQaOXX345qDDTpk0rbjg0nSEo1HWnCfFx4/2rX7P6juvqDoMhAIFIEmh97EezDJo0SFrSpH9znlqnZ+jQod9//31QhBT/EeTp6lNNv/5w+Zb8wm3/eVieqwuC8RCAAATsI9C6/KhbrMns5X/RRRc1GdSkZ2lp6Q8//HDhhRc2GeolT609quIUrCqqq6vTXDgvFY2yQAACEAgXgdbl5+mnn253Ztoi6OSTT9YipRs2bLj11lsTEhLOO++8dqfmlgtH981KSYzXvnM/bCkdkuvbIpYDAhCAAASCCNg79qPdGaQ3+tr0nHPOycnJmT9/fo8e3l+LMzkxfuyALIHOLywKws0pBCAAAQgYAq23fjpC6qWXXurI5e69dmJezvwft2v45/yDBri3FFgOAQhAwD4C9rZ+7LPb4Slr6WtZqI9PHW4n5kEAAhCIFgHkxxbyBw7MSoyP21Bcvq6ozJYMSBQCEICAywkgP7ZUYOfkxJF9fTMGaQDZwpdEIQAB9xNAfuyqw4bp1/S/2UWYdCEAAVcTQH7sqj6z9PUChn/sAky6EICAuwkgP3bV3wT/0js/btm1tbTCrjxIFwIQgIBrCSA/dlVdVufk4b1835wW0ACyizHpQgACLiaA/NhYefS/2QiXpCEAAZcTQH5srECz90/BKmYf2AiZpCEAAZcSQH5srDgjP99sLCkpr7IxG5KGAAQg4EICyI+NldYzI3VgTue6uk7a/NTGbEgaAhCAgAsJID/2VpoZ/uHjU3spkzoEIOBCAsiPvZVWP/zD5Dd7MZM6BCDgPgLIj711ZtYe/WrdjvKqGntzInUIQAACriKA/NhbXRr7yU1Pqaqp+2JN2/Ymt9csUocABCAQbQLIj701oM22mX5tL2JShwAE3EkA+bG93oz8MPvAdtBkAAEIuIoA8mN7dRn5WbymqKqm1vbMyAACEICASwggP7ZX1NDc9My0pLLKmmUbSmzPjAwgAAEIuIQA8mN7RcXHx5nVr1l71HbWZAABCLiHAPITibpi7dFIUCYPCEDAVQSQn0hUlxn+Wbh6e21tXSTyIw8IQAACjieA/ESiikb11ehPwo6yqpWbSyORH3lAAAIQcDwB5CcSVZSUEH/gwCzllF+4LRL5kQcEIAABxxNAfiJURRMH5Sin/FUsfR0h4GQDAQg4nADyE6EKmpCXrZzU+qnTBgwcEIAABGKeAPIToVtgbP/spIS4TSUVa7fvjlCWZAMBCEDAwQSQnwhVTlpywui+mcpsAcM/EUJONhCAgKMJID+Rq56Jeb7hn4JV2yOXJTlBAAIQcCoB5CdyNXNQXjdlxtqjkSNOThCAgIMJID+Rq5wDB2bHxXVata1sc0l55HIlJwhAAAKOJID8RK5a9Onpfr0ylF8+/W+Ro05OEICAQwkgPxGtGPb+iShuMoMABBxMIELyc/fdd2vfz+uvv97BKCJhGvITCcrkAQEIuIFAJOSnoKDg0UcfHTNmjBuA2GujWfp6+aadxWVV9uZE6hCAAAScTcB2+SktLb3gggsef/zx7GzfZ/8xfvRITxncvYvWPdDq1zGOguJDAAIxTsB2+Zk6depJJ5107LHHNge6oqKiJOBoLppn/Ol/80xVUhAIQKAjBOyVn5deemnx4sXTp09vwUSFZjYc/fv3byGmN4LYes4b9UgpIACBDhKwUX7Wrl173XXXzZw5MzU1tQUrp02bVtxw6JIWYnojyLR+lq4vLqus9kaJKAUEIACBdhBIbMc1IV6yaNGizZs3H3jggSZ+TU3NJ5988sgjj6i3LSEhwUokxX9Yp5539MtO652ZurG4/Is1OyYN6e758lJACEAAAk0SsLH1c8wxxyxZsuTLhmP8+PGag6CzQO1p0iZve2oCOsM/3q5iSgcBCIRCwMbWT3p6+qhRoywjunTpkpOTE+hjBcWaQ8M/b3y5gcXfYq3eKS8EIBBIwMbWT2A2uAMJmLVHv1hbVFldG+iPGwIQgEDsELCx9RME8aOPPgryidnTIbldszsnFZVVLVlfPG4gn0PF7I1AwSEQ0wRo/USh+jX8Y6Zfs/dPFOiTJQQg4AwCyE906oHZB9HhTq4QgIBjCCA/0akKIz9q/dTU1kXHAnKFAAQgEFUCyE908I/ondElOWFnefXyn3ZGxwJyhQAEIBBVAshPdPAnJsRr81PlnV+4LToWkCsEIACBqBJAfqKG30y/LlhVFDULyBgCEIBA9AggP1Fjb609WqcNGDggAAEIxBgB5CdqFb5//6zkhPitpRWrtpVFzQgyhgAEIBAlAshPlMB36pSalHBA/yxlz/BP1OqAjCEAgegRQH6ix75Tpwl5vtkHCwrZ+TSatUDeEIBAVAggP1HBXp/pxLwcuVj7IJp1QN4QgECUCCA/UQLvz/bAAVnxcZ3Wbt+9sXh3NO0gbwhAAAIRJ4D8RBx5QIbpqUkj+2TKg80XAqjghAAEYoIA8hPlajbTr5GfKFcD2UMAAhEngPxEHPneGVqLv+3tzRkEIAABjxNAfqJcwRMG+Sa/rdhUun1XZZRNIXsIQAACESSA/EQQdlNZ5XRN0e5zCmH+W1N48IMABDxLAPmJftXW97/x9U/0qwILIACByBFAfiLHurmcJg7qpqD8VXx82hwh/CEAAQ8SQH6iX6mm9bNsQ0lpRXX0rcECCEAAAhEhgPxEBHOLmfTJSuublaZtTxevZvOFFkkRCAEIeIgA8uOIymzY+4f+N0dUB0ZAAAIRIID8RABy61lMyPMN/7D2aOukiAEBCHiFAPLjiJo0wz9frt1RUV3jCIMwAgIQgIDNBJAfmwGHlvzg7l26d02urK79el1xaFcQCwIQgIC7CSA/jqi/uLg4Fn9zRE1gBAQgECkCyE+kSLeWj+l/Y+3R1jgRDgEIeIQA8uOUijStn0WrizQD2yk2YQcEIAAB2wggP7ahbWPC+/XOSE9J1Jen324saeOlRIcABCDgPgLIj1PqLCE+bpx/9WumXzulSrADAhCwkwDyYyfdNqbN2qNtBEZ0CEDAxQSQHwdVnln7QGuP1tUx/OOgesEUCEDADgL2ys+MGTPGjBmT4T8OOeSQ2bNn21EGz6Q5um9WSmK89p37YUupZwpFQSAAAQg0ScBe+enXr9/dd9+9aNGihQsX/uxnPzv11FOXLVvWpB14ikByYvzYAVly5Bey9ih3BAQg4HEC9srPySeffOKJJ+67775Dhw696667unbtOn/+fI8T7Vjx6vf+KdzWsWS4GgIQgIDTCSRGxsCamppXXnll165d6oILyrHCfxjPkpJYn3M8MS+nU6fvC1bR+gm6TTiFAAS8RsDe1o9oLVmyRI2elJSUK664YtasWSNGjAhCOH369MyGo3///kGhsXaqzjfNwF6/Y/e6orJYKzvlhQAEYoqA7fIzbNiwL7/8csGCBVdeeeXFF1/8zTffBPGdNm1accOxdu3aoNBYO+2Skjiqb6ZKXcDe27FW95QXAjFGwHb5SU5OHjJkyLhx49TK2X///R988MEgwmoYmalx5jcoNAZP66dfF7L1XAxWPkWGQAwRsF1+AlnW1tZqoCfQB3djAmbxN9Y+aEwGHwhAwEsE7J16oI61KVOmDBgwYOfOnS+88MJHH300Z84cL+GzoywT/Evv/Lhl19bSiu5dU+zIgjQhAAEIRJ2AvfKzefPmiy66aOPGjZpboO9PpT3HHXdc1MvscAOyOicP65m+fNPOgsLtU0b3dri1mAcBCECgfQTslZ8nn3yyfWbF+FVa/E3yo9V3kJ8YvxMoPgQ8TCCiYz8e5hjeok3I66YE2XouvFRJDQIQcBQB5MdR1VFvjFn7QBv/lJRXOdE+bIIABCDQYQLIT4cR2pBAr8zUgTmdteupNj+1IXmShAAEIBB9AshP9OugSQvM9Gv635qEgycEIOABAsiPQyuRreccWjGYBQEIhIkA8hMmkOFOxgz/fLVuR3lVTbjTJj0IQAAC0SeA/ES/Dpq0QGM/uekpVTV1X67d0WQEPCEAAQi4mgDy49Dqi4uLY/q1Q+sGsyAAgXAQQH7CQdGeNMzaoyx9bQ9dUoUABKJMAPmJcgW0kL2Z/Ka519U1tS1EIwgCEICAGwkgP86tNa38lpGaWFZZs2xDrG8C69xKwjIIQKC9BJCf9pKz/7r4+Dgz/Zqvf+yHTQ4QgECkCSA/kSbepvzY+6dNuIgMAQi4iADy4+jKMq2fhau312oFHg4IQAACHiKA/Di6Mkf1zUxLSthRVrVyc6mjDcU4CEAAAm0kgPy0EVhkoyclxB84MEt5au+fyOZMbhCAAATsJYD82Mu346mz9mjHGZICBCDgQALIjwMrZS+TrLVH6+oY/tmLDCcQgICrCSA/Tq++sf2zkxLifiopX7t9t9NtxT4IQAACIRNAfkJGFaWIackJo/tmKvMFhduiZALZQgACEAg/AeQn/EzDnuLEvBylyeJvYQdLghCAQBQJID9RhB9q1hPzshWVtQ9C5UU8CEDADQSQHxfU0riB3eLiOq3aVra5pNwF5mIiBCAAgRAIID8hQIp2FH16OrxXhqzg659oVwX5QwACYSOA/IQNpa0J1e/9U8jHp7ZiJnEIQCByBJCfyLHuSE7m658FyE9HIHItBCDgJALIj5Nqo3lbzNoHyzftLC6raj4WIRCAAARcQwD5cUdV9UhPGdy9i9Y90OrX7rAYKyEAAQi0SAD5aRGPkwJZ/M1JtYEtEIBARwkgPx0lGLHr63c+ZenriBEnIwhAwE4CyI+ddMOatpGfJeuKyyqrw5owiUEAAhCIAgHkJwrQ25dlv+y03pmp1bV1X67Z0b4UuAoCEICAcwggP86pi1YsiYuLY/p1K4wIhgAE3EPAXvmZPn36hAkT0tPTc3NzTzvttOXLl7uHjBMtNbMPWHvUiXWDTRCAQBsJ2Cs/H3/88dSpU+fPnz937tyqqqrJkyfv2rWrjRYSfQ8Bs/bB4jVFldW1e3xxQQACEHAhgURbbX733Xet9J955hm1gRYtWnTEEUdYnjjaRGBIbtfszklFZVVL1hePG+hbBpsDAhCAgEsJ2Nv6CYRSXFys027dugV6yl1RUVEScASFchpIQMM/9L8FAsENAQi4l0CE5Ke2tvb666+fNGnSqFGjgmBpfCiz4ejfv39QKKdBBMzsgze+3PD21xtXb9tVp4UQOCAAAQi4kEBcZJ5fV1555ezZs+fNm9evX78gSmr96DCeagVJgdROysjw7S/A0ZjANxtKTnzoU8s/PTVxZJ+MUX0yR/XVv4y87l0T4uOsUBwQgAAEokVAz3O1LFp4nkdCfq6++uo33njjk08+ycvLaxlEq+a2fHmMhL6zZOOnK7cu21D83cadlTV7zUFIS0rYr3e6T4r6ZI7sm7FvbnpyYoQauDECn2JCAAIhEmj1eW6v/Khpdc0118yaNeujjz7ad999WzW6VXNbTSGmIlTV1K7cVLp0Q7GaREvXF3+zsaSssiaQQHJC/LBeUqOMkf7m0fBe6alJCYERcEMAAhCwiUCrz3N75eeqq6564YUX1PQZNmyYKaHaYmlpac2VtlVzm7sQfxGoqa0r3LpLrSJJ0dL1JZKlneV7Lc+jfrkhPbqqVWQ660b0yeiaYu/UR+oFAhCIWQKtPs/tlR/N1ApC//TTT19yySVBntZpq+ZaMXG0SkBNz7Xbd0uE/ILkax5t21UZeJUqJy+ny0hfT12G+us0hpTVOTkwAm4IQAAC7SbQ6vPcXvlpq92tmtvWBIlvEZAabSqp8DWMfM2jEmnSxuJyK9Q4tKycb9DIqFHfjNz01KAInEIAAhAIkUCrz3PkJ0SSHoy2tbRi2QafDi3z99St3lYWVMjc9BT/LIYMXwupb2afzNTGzdmgSziFAAQgYAggP9wJoRIo3l2lKQz1Q0cbSn7YUhr0TZEWXPD30dU3jwZ26xzPJO9Q6RIPAjFHAPmJuSoPV4G1q9C3G6VGvkEjddat2LRTez0EJq5pC5q84J/F4Bs60l7giQlM8g4khBsCMU0A+Ynp6g9j4Suqa1b85Jvk7R89KpEyBS17mpoUv1/vejXSPO+hPfnkKIz4SQoC7iOA/LivzlxhsT45Uu+cGTTSr7rsdu39yVFSQpwUyLSNNHS0X6+MtGQ+OXJF3WIkBMJDAPkJD0dSaZlAbW3dqm27lmroqGFmnUaSAi/RIJGW6/ZNq/PP81avnb6ADYyAGwIQ8BgB5MdjFeqO4miS97qi3WZanTrrlqwv0Sy7INMH5XT2S5FvqTp11nXrwidHQYQ4hYC7CSA/7q4/z1i/uaTcfG8kNZIsrd+xO6hofbPSzPdGUiM1knIz+OQoiBCnEHAZAeTHZRUWI+Zu31Xp+97IP61Ov1orKKjgnZMTMtOS9C8jNSlDv2mJ1mm9wx/qi5CWqIVW+SApCCCnEIg6AeQn6lWAAa0T2Fnu++TIGjr6fnPp3nO8W0lB6iOV8kuRT6v8+pRofIyn79cfQY6uqZofHrwWVCsZEAwBCLSdQKvyw4qTbYfKFeEmoGkIBw3O0T+T8O7Kms07yzV5Qf9Kdlf7fst9bv9p/W9Juc9f/7TQalVNnZazC1rRrjkbtdKdvlgKFCTTtPLpU+c97S2/bvmipSQyYa85lvhDoEMEkJ8O4eNiOwhoivbAnC6hpKw5DtpjIlCZfBLlV6YSn3T5tWqPdFVLrbSUgxYC968FHjz+1GSO+p7JalpJjfy65VcvX7+faWn5m1Z+6eqSTDdgkxTxhEATBJCfJqDg5RYCGvLpkpKof32ymt3FI7As+nhWzSnTlqoXpwa5Ki4LaGNJscqqdlZUS6vKq2rLqyo27wyeuReYrOVWt15GamKgLBl30NiVPLVNrfZeSk2MZ6kIix6OWCOA/MRajcd0edWT1iNd/1JCoaCPmaRAlkr521W+5pTP4e8StHxMHLWs1BNYVFalf6Gkb+Jo4EqjUSmSoqR4zaHwaVJSvH5lqlqB0ifjoyB/nL2jJSak+uLUX2IuT0mMZy2+0PkTM4oEkJ8owidrRxPQQ9z0tvUPwUx1A6qdFDRG1aRQGa2yFomQaFXVVEvnQsgk1CjaYV265VMvXwOrXpwsYatXKYncXgrnj+xXPsX0x/Ern08a4yVpzC0MlT7xQiaA/ISMiogQaJ6Ans563Otfz9C+WFLTqqJa3Xo15dU10i3NtvA7airklqfvny9U7gpfHL/bF8d/ic+zIZr/8nL/5VIyY6CW49M/jYE1b2/bQjRfw5KxhnZYg3r5ZUwtsz2K5W/GSQLFRFMM9aOZhnLpXLnWn/r+Lx/j4Qvyx/TH8ce0Tv1RrAj1SSkZf5r+X39Sut6XibnW5/ad1J/63P6UGxwmsi93/1WBScnHnCqwvgANDsvHly5HOAggP+GgSBoQaCMBNa2MXLXxupaiq+vPKJalXhrr2l0ZIHJGzPzatkfkqn0K5z+t1zajfCYpMwNeY2CKoH+dOrWhX7ElW10eZvRMQiWJ0oCfJpyoQuXwn3ayTo1jT7T6CCZ+p4b4DZf7k/Klo2T9afoi+C5pSHBPBH92vtOAmAGhPn//qVLQITMsnz0p10eoj2nFr3f4StSpS3Jitp3LkSA/Lv87wHwINBDQs8ZMxGjw6Oj/1aOoFpXVLGvQtnqVkhqpEWZCfSLna4f5Gm1GzNT8UlusVknU+X5linHoV//5whoc/iiK44uhX/3nj+n/X0MKJin/JXvS9Kdaf2rS9Pv4E/Gl0UlNTH+qvqRMCvqf38d3bvm0A5Mu1xxKqbGvGN49Tt6/z8PnjbWvfMiPfWxJGQLuJqAX5+RE/fNNPXd3SVq03idOAWokp+/U57NHyepqfafy9AlqgyjKrRZnba1fiuTwn/p8jL/vt5NOdfi0qt7f5xMQIcBf6ltrnfou8Z/6cjTxA32USMNpg0O2+S+v99+Toy73B5kE/Vb5Yu459Wdk/M1V/lBlkWzzDl7IT4s3JoEQgIDXCfj6p/yjOQm+MSOOyBFge8rIsSYnCEAAAhCwCCA/FgocEIAABCAQOQLIT+RYkxMEIAABCFgEkB8LBQ4IQAACEIgcAeQncqzJCQIQgAAELALIj4UCBwQgAAEIRI4A8hM51uQEAQhAAAIWAeTHQoEDAhCAAAQiRwD5iRxrcoIABCAAAYsA8mOhwAEBCEAAApEjgPxEjjU5QQACEICARQD5sVDggAAEIACByBFAfiLHmpwgAAEIQMAigPxYKHBAAAIQgEDkCDhrwwX/VhqdSkpKIgeAnCAAAQhAwAYC5klunupNJu8s+dm5c6es7N+/f5O24gkBCEAAAu4ioKd6ZmZmkzbHtSBNTV5gq2dtbe2GDRvS09Prt39qV2aSXAnY2rVrMzIy2pWA0y+igE6vodbsowZbI+T0cM/XoCqg42WUuEh7+vTpEx/f9CiPs1o/srJfv35hufWkPV6VH8OHAoblPoliItRgFOGHJWvP16AodbCMzbV7DP+mRSksdUMiEIAABCAAgeYIID/NkcEfAhCAAARsJJBw22232Zh8lJJOSEg46qijEhOd1bUYRhgUMIwwo5IUNRgV7GHM1PM1KFZ2l9FZUw/CeHOQFAQgAAEIOJkAnW9Orh1sgwAEIOBZAsiPZ6uWgkEAAhBwMgHkx8m1g20QgAAEPEsA+fFs1VIwCEAAAk4m4DX5+e///u9BgwalpqYedNBB+fn5TkYfim2al6gFIKxj+PDh5qry8vKpU6fm5OR07dr1zDPP3LRpUyipOSTOJ598cvLJJ+tbaJXr9ddft6zSN9K33HJL796909LSjj322JUrV1pB27dvv+CCC/QFXFZW1n/+53+WlpZaQc50NFfGSy65xKpNOU444QTLfheVcfr06RMmTNDqJLm5uaeddtry5cutqp/N/AAACnNJREFUUrRwZ65Zs+akk07q3Lmzrvr9739fXV1tXeU0RwsF1JTawBq84oorLONdVEDZPGPGjDFjxpivSg855JDZs2ebgkS0BvU375njpZdeSk5Ofuqpp5YtW/arX/1Kjyo9l11dultvvXXkyJEbG44tW7aY4uim18JCH3zwwcKFCw8++OBDDz3URcV85513brrpptdee023+6xZsyzL7777bn0jLUH66quvTjnllLy8vN27d5tQPab333//+fPnf/rpp0OGDDnvvPOsq5zpaK6MF198scrSUJ8bJTmW/S4q4/HHH//0008vXbr0yy+/PPHEEwcMGKAXAlOQ5u5Mic2oUaP0VvHFF18ITvfu3adNm2aV3WmOFgp45JFH6tli1WBxcbEx3l0FlM1vvvnm22+/vWLFCr09/PGPf0xKSlKFyj+SNdjJaRXfEXsmTpyoNoFJoaamRu/XeovpSIJRv1byo8dukBk7duzQvfLKK68Y/2+//VbP8c8//zwomvNPA+VHy/316tXrL3/5izFbZUxJSXnxxRd1+s033yhmQUGBCdJrml4/169f7/wCysLAMupU8nPqqac2tty9Zdy8ebPK+PHHH6tQLdyZkhwtqfXTTz+ZsuvVW+/dFRUVjVE4zSewgLJN8nPdddc1NtK9BTRlyc7OfuKJJyJcg97pfKusrFy0aJFer/THoEP3utx6KJtT9/6qD0o6OnjwYPU+qXWvgqiYVVVVVknVI6fXT7eXtLCwUM8mq1BqBqn71BRKv2rIjh8/3lSi4qhyFyxY4NI6/eijj9T7NGzYsCuvvHLbtm2mFO4to17/VYRu3brpt4U7UwUcPXp0z549TXnVvNCKluqlMKdO/g0soLFz5syZar2pMacGXFlZmfF0bwH1pq5+o127dqkLLsI16J11AbZu3SqO1v2te0Lu7777zsl3dqu26RH8zDPP6FGlxv7tt99++OGHq4Gsx7T6GPVEti5XSeVpnbrRYewPqj7jqV89r61CaTELPexcWl71sJ1xxhnqV/zhhx/U4zFlyhQ9tvRtuUvLqDbr9ddfP2nSJD2LVUEqRXN3poKCKtfEt6rVmY6gAsrI888/f+DAgXoj/Prrr//whz+o58p0I7uxgEuWLJHkaLBHQ8jqBh8xYoR6UyNZg96RH2fevh20So8nk4IGCSVFuu9ffvlljcx3MFkujxaBc88912StpoDqdJ999lFj6JhjjomWPR3MV33deh+aN29eB9Nx7OWNC3j55Zcba1WDmiajutObhOrRsUVowTC910pv1Lx79dVX1S2sHtQWItsR5J3ONzWH9RYZOAdMbg0n2EEtKmmquTN06NDvv/9ehVJPo3ppLTM8UFJTU01Wn4JM/7spr8Z4NWLvgZpVh6puWlWoyuXGMl599dVvvfXWhx9+aG2SolI0d2cqKKhyTamte9iBjsYFDDJSb4TysWrQdQVUQ0cTecaNG6cxco0xP/jggxGuQe/Ij1CKoyaDmVtErWa51bQMumPce6rJRXrP0guXiqmpB1ZJ1fzXmJDbS6r+KN36VqE0MKDRHVMo/Upr1Stt6u5f//qXKtf85bu3NmX5unXrNPajCpXbXWXUYLUezequUV2o4qxaaOHOVAHV1WO9RsydO1dTD9TbY13rKEdzBQwyUk0H+Vg16KICBhVEp/qb0kyQSNdg4ykc7vXRAJqmS2mwRPOI1EZWc0Edsu4tjiz/3e9+p84ZDcv/+9//1pC7Xpb1Byx/TY7UdAP98Wvitf6wdbiomNoAUbNvdeimv+++++RYvXq17NfEa1XZG2+8oV51TQ8Lmng9duxYCZL6efbdd1/nT7xusozyvOGGGzTYowp9//33DzzwQJVFPe+m7jQs5JYyatKE5obozrTmH2sE3pSiuTvTzEuePHmyHtnvvvtujx49nDzxurkCqqFzxx136I9ONagbVe3XI444whTcXQWUzTfeeKN621QQ/bnJrdmk7733nvwjWYOemngtdg8//LCey2oJaRK2PhMxd4Z7f3/+85/r3UrF6du3r9y6+01Z9EHMVVddpbmS+ojv9NNP11PARWVUd03Qy5f6nWW/3r9uvvlmDVDrHUJd6mrVWYVSK0GSowFSvTJfeumleo5bQc50NFlGPaP1/NWTV41XDePp85HA1yMXlTGo+nSqz4BMRbRwZ65atUpjmRq51FuU3qs0e9OZdSermiuguhmkN5r5oltU3Vb6eNb67kdXuaiAsvayyy7TTahni25I/bkZ7ZF/JGuQDRca32n4QAACEICA7QS8M/ZjOyoygAAEIACB8BFAfsLHkpQgAAEIQCBkAshPyKiICAEIQAAC4SOA/ISPJSlBAAIQgEDIBJCfkFEREQIQgAAEwkcA+QkfS1KCAAQgAIGQCSA/IaMiIgQgAAEIhI8A8hM+lqQEgXAT0LfogRvChjt50oNANAkgP9GkT96uJhC0c7akQgvnuLpEGA+BSBJgw4VI0iYvrxGQ3mi9GatUWovFcuOAAARaJkDrp2U+hEKgJQLSGy3UbR1ahU+x1QzSZtJmfTOtSqnNVKwktCjyz372M617lpOTo1VxtYq5FfTUU0+NHDlSCWqVP60nbflrH0Ut66fF/bQ+6Ztvvmn544CA2wkgP26vQex3IgGtnXrmmWd+9dVX2iJdW8x9++23slL7GWuTaUlUQUHBK6+8okWvLZmRXGlnMwmS9Ekao+UsrVJpl9tzzjlHyxKfeOKJSk17HVlBOCDgbgJa4pQDAhBoBwEt1K0dDrsEHHfddZfS0RNBq9ZbCWprIi3gr9PHHntM2qMWjwl6++234+PjzaLX2rz5pptusi6xHErqT3/6kzk1TaXZs2dboTgg4GoCjP24++0B66NL4Oijj1bDxbJBS/Ebt3ZgsjzlNvuSqQ2kPSWlViZo0qRJ2mNC+0qos27Dhg3NbbmtPblNfF2o/SasHdus9HFAwKUEkB+XVhxmO4KAJCGwo6x9NmkoqIULtTmQFSqhkmJZpzgg4GoCjP24uvow3qEEtNWhZZnc++23n071q9EgjQCZIO1gq863YcOGpaenDxo0yNpo3LoQBwS8TYDWj7frl9LZS6CiokKDN1YeiYmJ2spTp5pZMH78+MMOO2zmzJn5+flPPvmkPDVx4NZbb9WI0W233bZly5Zrrrnmwgsv1O6uCpKPhotyc3M1X057uUqZFGoliwMCniSA/HiyWilUhAi8++67midtZaamzHfffadTTVd76aWXtCG6Ql988cURI0bIU5On58yZc911102YMEFuTY277777zLXSpPLy8vvvv/+GG26QgJ111llWmjgg4FUCbLbt1ZqlXFEjoBGaWbNmnXbaaVGzgIwh4AYCjP24oZawEQIQgIDnCCA/nqtSCgQBCEDADQTofHNDLWEjBCAAAc8RoPXjuSqlQBCAAATcQAD5cUMtYSMEIAABzxFAfjxXpRQIAhCAgBsIID9uqCVshAAEIOA5AsiP56qUAkEAAhBwAwHkxw21hI0QgAAEPEcA+fFclVIgCEAAAm4ggPy4oZawEQIQgIDnCCA/nqtSCgQBCEDADQT+P763P6T048VSAAAAAElFTkSuQmCC)"],"metadata":{"id":"Z2lHiH6OliYg"}},{"cell_type":"markdown","source":["We examined the effect of using different learning rates with our debugging dataset for 300 epochs. The results graphed below show that the learning rate we used (0.01) was optimal\n","\n","![image.png](data:image/png;base64,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)"],"metadata":{"id":"YBGs_PDclxmg"}},{"cell_type":"markdown","source":["We tried our model with dropout layers with varying dropout rates. But since our model is not overfitting to our training data, dropout only slightly negatively impacted our final performance. Thus we removed those layers."],"metadata":{"id":"KK5hcfrstmCc"}},{"cell_type":"markdown","source":["## Trying normal neural network for training\n","We tried a simpler model without a GNN on our debugging dataset. The end results were similar to the model above.\n"],"metadata":{"id":"uFr2KFpKqam_"}},{"cell_type":"code","source":["import pandas as pd\n","import torch\n","import torch.nn as nn\n","import torch.optim as optim\n","from sklearn.preprocessing import LabelEncoder\n","from sklearn.preprocessing import OneHotEncoder\n","from sklearn.preprocessing import StandardScaler\n","from sklearn.model_selection import train_test_split"],"metadata":{"id":"PYvtQQXyqrAJ"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["import pickle\n","import pandas as pd\n","\n","# NOTE FOR GRADERS - please add the folder linked below to your drive and alter this string to match the location of the \"pickled_data\" folder:\n","# https://drive.google.com/drive/folders/1upZ3Oh242b83WTBUQB4HS64M0BlY2PHE?usp=drive_link\n","shared_folder_path = f'/content/drive/MyDrive/CS547 Project Group 11/pickled_data'\n","save_path = f'{shared_folder_path}/debugging_dataset_2023.pkl'"],"metadata":{"id":"W53yB6v9twXt"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["debugging_df = pd.read_pickle(save_path)\n","debugging_df = debugging_df.drop(columns=['start_station_id', 'end_station_id', 'ride_id'])\n","debugging_df"],"metadata":{"id":"np6dQzIJquKs"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["### Import weather data per date"],"metadata":{"id":"RxwASMxBrVvk"}},{"cell_type":"code","source":["weather_path = f'{shared_folder_path}/weather_data.csv'\n","weather_df = pd.read_csv(weather_path)\n","weather_df['date'] = pd.to_datetime(weather_df['DATE'])"],"metadata":{"id":"XNFU9kUxqz2i"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["weather_df = weather_df.drop(columns=['STATION', 'NAME', 'DATE'])\n","weather_df = weather_df.fillna(weather_df.mean())\n","weather_df"],"metadata":{"id":"7MbFL4pnq2qm"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["### Encode the categorical data using one hot **encoding**"],"metadata":{"id":"3GH0mmIvryMA"}},{"cell_type":"code","source":["# Encode 'rideable_type' and store the mapping\n","debugging_df['rideable_type'], rideable_type_mapping = pd.factorize(debugging_df['rideable_type'])\n","rideable_type_mapping = dict(enumerate(rideable_type_mapping))\n","\n","# Encode 'member_casual' and store the mapping\n","debugging_df['member_casual'], member_casual_mapping = pd.factorize(debugging_df['member_casual'])\n","member_casual_mapping = dict(enumerate(member_casual_mapping))\n","\n","# Combine unique values for consistent encoding\n","unique_regions = pd.unique(debugging_df[['start_region', 'end_region']].values.ravel())\n","region_mapping = {region: code for code, region in enumerate(unique_regions)}\n","\n","# Map the codes using the consistent mapping\n","debugging_df['start_region_ID'] = debugging_df['start_region'].map(region_mapping)\n","debugging_df['end_region_ID'] = debugging_df['end_region'].map(region_mapping)\n","\n","# Similarly for start_station_name and end_station_name\n","unique_stations = pd.unique(debugging_df[['start_station_name', 'end_station_name']].values.ravel())\n","station_mapping = {station: code for code, station in enumerate(unique_stations)}\n","\n","# Map the codes using the consistent mapping\n","debugging_df['start_station_ID'] = debugging_df['start_station_name'].map(station_mapping)\n","debugging_df['end_station_name_ID'] = debugging_df['end_station_name'].map(station_mapping)"],"metadata":{"id":"RmLisg-Eq4PQ"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["### Getting more features for the date"],"metadata":{"id":"aV8HkQ-0sEmZ"}},{"cell_type":"code","source":["debugging_df['date'] = pd.to_datetime(debugging_df['started_at']).dt.date\n","debugging_df['date'] = pd.to_datetime(debugging_df['date'])\n","debugging_df['start_hour'] = debugging_df['started_at'].dt.hour\n","debugging_df['start_day'] = debugging_df['started_at'].dt.day\n","debugging_df['start_month'] = debugging_df['started_at'].dt.month\n","debugging_df['start_weekday'] = debugging_df['started_at'].dt.weekday"],"metadata":{"id":"UGF_4UnJq5CY"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Convert 'date' columns to datetime in both DataFrames\n","debugging_df['date'] = pd.to_datetime(debugging_df['date'], errors='coerce')\n","weather_df['date'] = pd.to_datetime(weather_df['date'], errors='coerce')\n","\n","# Verify the date columns\n","print(debugging_df['date'].head())\n","print(weather_df['date'].head())"],"metadata":{"id":"K1_l5lieq69Q"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["### Joining the data with the weather data on date"],"metadata":{"id":"rJllqOVtsNxB"}},{"cell_type":"code","source":["debugging_df = pd.merge(debugging_df, weather_df, on='date', how='left')"],"metadata":{"id":"QeYxSBgQq84A"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["debugging_df.drop(columns=['start_station_name', 'end_station_name', 'start_region', 'end_region', 'ended_at', 'end_lat', 'end_lng', 'end_station_name_ID', 'started_at', 'date'], inplace=True)"],"metadata":{"id":"OC5hv6Zjq-y_"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["### Normalize the longitude and latitude"],"metadata":{"id":"w0TgkYzAsZpL"}},{"cell_type":"code","source":["# Normalize numerical features\n","scaler = StandardScaler()\n","numerical_features = [\"start_lat\", \"start_lng\"]\n","debugging_df[numerical_features] = scaler.fit_transform(debugging_df[numerical_features])"],"metadata":{"id":"L-rdc1ZFrAFJ"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["### Predict the end region"],"metadata":{"id":"9qSwv4ejsdlH"}},{"cell_type":"code","source":["label_name = 'end_region_ID'"],"metadata":{"id":"6OfoXoYMrBux"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Verify the new encoding\n","num_classes = len(debugging_df[label_name].unique())\n","print(f\"Number of classes after encoding: {num_classes}\")\n","print(f\"Maximum class index after encoding: {debugging_df[label_name].max()}\")"],"metadata":{"id":"6MnK70IErDK9"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Create and fit the label encoder\n","label_encoder = LabelEncoder()\n","debugging_df[label_name] = label_encoder.fit_transform(debugging_df[label_name])"],"metadata":{"id":"VQmSsOLdrE8X"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Separate features and target\n","features = debugging_df.drop(columns=[label_name])\n","target = debugging_df[label_name]"],"metadata":{"id":"nCs_DCB7rF3E"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["### Create train and test sets"],"metadata":{"id":"f2i8r2CwsjFj"}},{"cell_type":"code","source":["X_train, X_test, y_train, y_test = train_test_split(features, target, test_size=0.2, random_state=42)\n","\n","# Convert to PyTorch tensors\n","X_train = torch.tensor(X_train.values, dtype=torch.float32, device=DEVICE)\n","X_test = torch.tensor(X_test.values, dtype=torch.float32, device=DEVICE)\n","y_train = torch.tensor(y_train.values, dtype=torch.long, device=DEVICE)\n","y_test = torch.tensor(y_test.values, dtype=torch.long, device=DEVICE)"],"metadata":{"id":"uATK6zSdrG9x"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["### Remove NaN values for training"],"metadata":{"id":"TwZmc-v8smmj"}},{"cell_type":"code","source":["# Check for NaN or infinite values in the input data\n","print(torch.isnan(X_train).any())\n","print(torch.isinf(X_train).any())\n","print(torch.isnan(y_train).any())\n","print(torch.isinf(y_train).any())\n","\n","# If any NaN or inf values are present, you can remove or replace them\n","X_train = X_train[~torch.isnan(X_train).any(dim=1)]\n","y_train = y_train[~torch.isnan(y_train)]\n"],"metadata":{"id":"C6BcttKurId4"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["class RidePredictionModel(nn.Module):\n","    def __init__(self, input_dim, num_classes):\n","        super(RidePredictionModel, self).__init__()\n","        self.fc1 = nn.Linear(input_dim, 256)\n","        self.fc2 = nn.Linear(256, 128)\n","        self.fc3 = nn.Linear(128, num_classes)\n","        self.relu = nn.ReLU()\n","        self.dropout = nn.Dropout(0.3)\n","\n","    def forward(self, x):\n","        x = self.relu(self.fc1(x))\n","        x = self.dropout(x)\n","        x = self.relu(self.fc2(x))\n","        x = self.fc3(x)\n","        return x"],"metadata":{"id":"ROmAkhnwrLfZ"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Define model\n","input_dim = X_train.shape[1]\n","num_classes = len(debugging_df[label_name].unique())\n","model = RidePredictionModel(input_dim, num_classes)\n","model.to(DEVICE)\n","\n","# Loss and optimizer\n","criterion = nn.CrossEntropyLoss()\n","optimizer = optim.Adam(model.parameters(), lr=0.001)"],"metadata":{"id":"EQMNp3RkrNOP"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["import matplotlib.pyplot as plt\n","\n","# Initialize a list to store loss values\n","loss_values = []\n","\n","# Training loop\n","num_epochs = 40\n","batch_size = 512\n","\n","for epoch in range(num_epochs):\n","    model.train()\n","    epoch_loss = 0\n","    permutation = torch.randperm(X_train.size(0))\n","\n","    for i in range(0, X_train.size(0), batch_size):\n","        indices = permutation[i:i + batch_size]\n","        batch_X, batch_y = X_train[indices], y_train[indices]\n","\n","        # Forward pass\n","        outputs = model(batch_X)\n","        loss = criterion(outputs, batch_y)\n","\n","        # Backward pass\n","        optimizer.zero_grad()\n","        loss.backward()\n","        optimizer.step()\n","\n","        # Accumulate loss for the epoch\n","        epoch_loss += loss.item()\n","\n","    # Store the average loss for this epoch\n","    average_loss = epoch_loss / (X_train.size(0) // batch_size)\n","    loss_values.append(average_loss)\n","\n","    print(f\"Epoch [{epoch + 1}/{num_epochs}], Loss: {average_loss:.4f}\")"],"metadata":{"id":"yVt_zCwnrQEt"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Plot the loss curve\n","plt.figure(figsize=(8, 5))\n","plt.plot(range(1, num_epochs + 1), loss_values, marker='o', label='Training Loss')\n","plt.xlabel('Epoch')\n","plt.ylabel('Loss')\n","plt.title('Loss Curve')\n","plt.legend()\n","plt.grid()\n","plt.show()"],"metadata":{"id":"fLu1hsJVrR3Q"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Evaluate the model\n","model.eval()\n","with torch.no_grad():\n","    y_pred = model(X_test)\n","    predicted_classes = torch.argmax(y_pred, axis=1)\n","    accuracy = (predicted_classes == y_test).float().mean()\n","\n","print(f\"Test Accuracy: {accuracy:.4f}\")"],"metadata":{"id":"i7yoinFdrTaJ"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["# Sources\n","- Course slides and example code for Feed Forward Neural Nets\n","- [Explanation of Graph Neural Networks](https://lightning.ai/docs/pytorch/stable/notebooks/course_UvA-DL/06-graph-neural-networks.html#)\n","- ChatGPT and Perplexity to generate some utility code snippets (cited where used)\n"],"metadata":{"id":"048_cEj5aQQW"}},{"cell_type":"markdown","source":["https://lightning.ai/docs/pytorch/stable/notebooks/course_UvA-DL/06-graph-neural-networks.html#"],"metadata":{"id":"76AHDsioipjq"}},{"cell_type":"code","source":[],"metadata":{"id":"ZG9bgQg2iq-1"},"execution_count":null,"outputs":[]}]}