Please watch the following videos to learn about Gamer Tweet Detector: https://www.youtube.com/watch?v=S2zvvUyJm7I&list=PLieKBD_gMWPtRjVROAcFybNaToxvJ2eW9
We wanted a way to easily determine if a tweet was a gamer tweet or not.
What it does
This application takes in a tweet url and uses deep learning to calculate a Gamer Tweet Percentage (GTP) for the given tweet and output it to the user.
How We built it
This backend was built in python, using the
Twython library to access the Twitter API and analyze tweets. The
sklearn library to perform logistic regression on our dataset. Our frontend was built using the Flask and Node frameworks
Challenges We ran into
Understanding the machine learning process and how to curate an acceptable dataset, as well as correctly interfacing that with the frotend
Accomplishments that We're proud of
being able to calculate an f1 score with our dataset
What We learned
Machine Learning Stuff
What's next for Gamer Tweet Detector
Clean up data curation scripts, and further improve our data model.