WhatPy

Machine Learning and Data Mining with Whatsapp Chat logs

Overview:

This is a project to learn various NLP and data mining techniques while doing something fun and interesting. We will explore different language feature extraction methods, and a variety of ML classification or clustering methods. Through this project, we hope to become familiar with the basics of NLP techniques as well as gain exposure to a variety of NLP and ML toolkits for python.

Breakdown of Project:

1. Parse whatsapp chat logs
2. Tagger + feature extraction
3. ML supervised and/or unsupervised learning with various combinations of data
4. Analysis & Comparison between models
5. Extra: IBM BlueMix Speaker Classification

Acknowledgements:

gtadiparthi's Whatsapp parser was tweaked to fit current whatsapp chatlog format