All Projects


Alt text


TwitOff is a fun and interactive application that allows you to predict which user is more likely to post a particular tweet from their account. Build using Flask, the script pulls data from the Twitter Developer API and populates a SQLite datebase. Afterward, the user inputs a string into a prompt and a logistic regression is run on our data with accuracy used as our metric. Finally, to make our application stand out a NES.css template is used and the application is deploted to Heroku.

Image 1

Model and Metric

TwitOff uses a Logistic Regression algorithm to predict the probabilty of a given tweet with a 72.1% accuracy score. For more information on Logistic Regression and how it works please read Retweet Predictive Model in Twitter