Using the politwoops dataset, we use machine learning to analyze the tweets, breaking the project up into the following 4 categories:
- exploratory analysis and text-mining (notebook)
- deletion times (how long tweets were live) (notebook)
- sentiment analysis (notebook)
- hashtag recommendation system (notebook)
for more info on the open-source Politwoops tweet collection method visit their github. the web application used to display the tweets on https://projects.propublica.org/politwoops/ is also open-source
Shayna Anderson-Hill contributed frequently to this project