A tool to explore FiveThirtyEight's Russian Troll Tweets dataset
Branch: master
Clone or download
Latest commit 82e11f2 Aug 3, 2018
Type Name Latest commit message Commit time
Failed to load latest commit information.
models First commit. Aug 2, 2018
model-topics.ipynb Fixing a lot of methodology stuff. Aug 3, 2018


Russian Troll Tweets Exploration

What is this?

The good people at FiveThirtyEight were kind enough to release the tweets and handles of Russian Trolls in the lead-up and following the 2016 election. These data were assembled by Clemson University researchers Darren Linvill, & Patrick Warren. I'll let FiveThirtyEight explain more, here. I'm just exploring the data they released here.


  • Run git clone git@github.com:fivethirtyeight/russian-troll-tweets.git to get the Russian tweets.
  • Run python3 -m venv env to set up your virtual environment.
  • Run source env/bin/activate to activate your virtual environment.
  • Run pip install -r requirements.txt to install the Python packages you'll need.
  • Run npm install -g http-server (you might want to install NVM first).


  • Run jupyter notebook and pull up model-topics.ipynb.
  • Run everything in there.

Getting visualizations working

  • Run cp intermediate-data/english-only-doc2vec-data.json doc2vec-graphics/som-account-type-viz/data/data.json.
  • Run http-server from within the doc2vec-graphics/som-account-type-viz/ folder.
  • Go to http://localhost:8080.

When you're done

  • Run deactivate.