/Sephora-ML-Project

End-to-end machine learning project to predict number of loves of Sephora products.

Primary LanguageJupyter Notebook

End-to-End Machine Learning Project with Sephora Data

Problem Statement:

How does the popularity of a skincare product change with the product's information (skin concern, skin type, price, etc)?

We will use Sephora's number of loves to quantify popularity, and will address this problem statement with regression models using supervised learning. The goal is to find out the features that makes a product popular from building models that can predict a product's popularity.

Business Use Cases:

For skincare manufacturers, understanding what makes a product popular is important, as they can use their resources to produce products which will be popular. It also help with their pricing strategy. For example, are products made for dry skin more popular than those for oily skin? Or are moisturisers more well-loved than treatments?

For consumers, we can build interactive visualization tools from the a large dataset containing all products, which can help consumers with comparing the reviews of different products.

Packages Used:

  • Pandas
  • Numpy
  • Selenium
  • Beautiful Soup
  • Matplotlib
  • Seaborn
  • SciPy
  • Scikit-learn

Machine Learning Models Used:

  • Multiple Linear
  • Lasso
  • Nearest Neighbor
  • Decision Tree
  • Random Forest
  • Gradient Boosting Regressor

Relevant Links: