/Cosmetics-Chemical-Composition-Analysis-and-Visualization

Choosing cosmetics can be tough and risky. Instead of guessing, let's use data science to predict suitable products. This notebook creates a content-based recommendation system by processing ingredient lists of 1,472 Sephora cosmetics with word embedding, and visualizes similarities using t-SNE and Bokeh.

Primary LanguageJupyter Notebook

No issues in this repository yet.