Ecommerce Recommendation Engine

Objective:

Engineer a product recommendation system for an e-commerce website to increase customer retention and sales..

Featured Techniques:

• Feature Engineering
• Collaborative Filtering
• Surprise Package
• SVD Dimensionality Reduction

Methodology:

Users are separated into repeat customers and first time customers and the recommendation system works as follows.

• Repeat Customers
  • Collaborative filtering recommendation
  • Hot Products
  • Popular in your area
• New Customers
  • Hot products
  • Popular in your area

Data Source

https://www.kaggle.com/olistbr/brazilian-ecommerce#olist_products_dataset.csv

Folder Structure

Clean code for implementation is found in python files. Exploratory data analysis and model selection and tuning are found in jupyter notebooks within the notebooks folder.