A web application that takes a text string, and determines who is most likely to tweet the text by comparing the tweets of two Twitter users. This application was created using the Python library Flask-SQLAlchemy. It collects tweets from the Twitter API, stores the data in a PostgreSQL database, and is deployed on Heroku. The tweets are converted to embedding vectors using Basilica.ai, which is then used to run a logistic regression. The model is then used to predict the probability of a text string belonging to one of two Twitter users, and outputs the username with the highest probability.
Application link: https://jldaniel77-twitoff.herokuapp.com/