This project attempted to improve product recommendation in e-commerce by developing a recommendation system that applied collaborative filtering algorithm using Twitter data as a basis for making recommendations to other members of an e-commerce store. The recommendation system was built using Python 3.8. The major modules in the recommendation system includes the Data Extraction Module, the Network Module, and The Conjunction module (which combines the metrics from the network and the metrics from the website database). The python packages that played a vital role in the conclusion of this project include, Tweepy, Networkx, Pandas, intertools, matplotlib, re, scipy.spatial.distance and csv. The project provided interesting recommendations to users based on their closeness with other Twitter users registered on our e-commerce store. It also eradicated the problem of cold-start finish, which has been a persistent major problem in recommender systems.
