Twitter Auto-completer for Members of Congress
This project is a tweet auto-completer for members of Congress. I used the Twitter API & the DocNow hydrator to create a custom dataset. I then used the GenSim library to generate a custom word2vec representation and finally used a Keras LSTM model to auto-complete tweets.
I used subsets of the following datasets from George Washington University stored on Harvard Dataverse:
- 115th US Congress Tweet Ids
- U.S. Government Tweet Ids
- 2016 United States Presidential Election Tweet Ids