Exploratory Data Analysis (EDA) in R

The EDA process for the Plant Ecosystem Analysis project involved a comprehensive examination of the "PlantData.txt" dataset using R. This dataset contains 137 observations of 9 variables related to various geographical and environmental gradients impacting plant ecosystems. The primary objective was to explore the relationship between Native plant species Richness (NR) and ecological/environmental factors such as Area, Latitude, Elevation, Distance to significant landmarks (Dist), Soil types, Age of ecosystems (Years), Time since last deglaciation (Deglac), and adjacent Human population densities (Human.pop).

Initial Data Loading and Inspection

• Data Import: The dataset was loaded into R using the read.csv function, specifying the path to the "PlantData.txt" file.
• Initial Peeking: Utilized the head() function to examine the first few rows of the dataset, ensuring proper loading and to get an initial sense of the data structure.

Data Structure and Summary

• Structure Inspection: The str() function revealed the dataset as a data frame with 137 observations and 9 variables, with all variables being numerical.
• Summary Statistics: Employed summary() to obtain a descriptive summary of each variable, including measures of central tendency and dispersion.

Missing Values and Unique Counts

• Checked for missing values using colSums(is.na(Data)) and confirmed the dataset had no missing values.
• Calculated unique value counts for each column to determine the diversity of data and assess if any categorical variables were present.

Pairwise Relationships and Correlations

• Used the GGally package to generate pairwise relationship plots with ggpairs(), aiding in visualizing correlations and potential relationships between variables.
• Constructed a correlation matrix to quantify the strength of relationships between variables, identifying particularly strong correlations worth further investigation.

Distribution Analysis

• Visualized the distribution of each numerical variable using histograms, noting the distribution shape, skewness, and potential outliers.
• Generated boxplots for each variable to further assess data spread and identify outliers.

Correlation Plot

• Created a correlation plot using the corrplot package, visually representing the correlation matrix and highlighting significant correlations.

Transformations and Diagnostics

• Applied transformations to certain variables (e.g., log transformation to Human.pop) to address skewness and improve model fitting.
• Conducted diagnostics using plots of residuals, Q-Q plots, and other checks to evaluate model assumptions and fit.

Model Fitting and Selection

• Explored multiple linear regression models, starting with a full model including all predictors and progressively refining the model based on statistical significance and model diagnostics.
• Utilized backward selection, AIC, BIC, and adjusted R-squared values to identify the most parsimonious and informative model.

Final Model and Interpretation

• The selected final model incorporated log-transformed predictors and demonstrated the best fit based on adjusted R-squared and residual diagnostics.
• Key predictors identified as significantly impacting Native plant species Richness included Area, Elevation, and Soil types.

This detailed EDA process laid a solid foundation for subsequent analysis and modeling, providing valuable insights into the complex relationships within the plant ecosystem data.