Big Data Analytics Project: Impact of Social Signal on E-Commerce Search Engine

Description:

    We had developed this project for Unbxd Inc., basically the idea is to read and analyze business specific Social Signals trending on social Media like Twitter. These Signals gets used to calculate the product score. Finally, based on the product score and product ranking, relevant products (which is in trend) gets displayed on www.express.com in the very first page.

    Overall the project aims to improve user search on www.express.com by improving product rank using Social Signal score. As a result, products with higher social signal score (trending products) gets higher conversion rate in less advertisement cost.

Client: Unbxd Inc. and www.express.com

Language: Python

Libraries Used: Tweepy, Keras, Sklearn, Word2Vec, Glove, Pandas

Models for sentiment analysis/prediction: Logistic Regression, Random Forest, Neural Network

Models for Product identification/recognition: Logistic Regression, SVM, Naïve Bayes, KNN, Neural Network.

Challenges: Manual labelling of training data for modelling,

Working as a C++ developer on  Data Protector, a backup and recovery product.

Worked as a c++ developer on a Financial project at paypal.

Worked as a c++ developer at TIVO set top box application