This project compares 3 dimensionality reduction techniques (linear: PCA & non-linear: t-SNE, UMAP) to produce a similarity map. The map draws similarities among occupations skills and knowledge features, using ONET occupational database. The results can be visualized here:
- PCA: https://plot.ly/~fanny_kassapian/3/
- t-SNE: https://plot.ly/~fanny_kassapian/5/
- UMAP: https://plot.ly/~fanny_kassapian/1/
I explain how I use dimensionality reduction to build a non-geographical map here: https://towardsdatascience.com/how-to-build-a-non-geographical-map-1-8d3373e83d6c