QUINN DIZON

DATA SCIENTIST, DEVELOPER, MUSICIAN

Projects

Stack Overflow Tag Predictor

2019

A web app that uses natural language processing, text vectorization, and machine learning models to predict the primary language tag of a Stack Overflow post.


Technologies

  • Python
  • Scikit-Learn
  • Flask + Heroku
  • React + Netlify

Features

  • Uses Natural Language Processing (NLP) to determine the topic from raw post text.
  • Achieved ~82% accuracy over five language tags (python, javascript, java, C#, C++)
  • The best performing model (logistic regression) is deployed to a Flask backend API with a React frontend UI. Backend deployed to heroku, front end deployed to netlify.