Check political bots on social networks
The goal is to show information about politicians regarding who is retweeting their Twitter tweets, if it is humans or bots.
We use information from Truthy BotOrNot, a project from Indiana University, which calculates a probability of a Twitter account being a bot or a human. We check that on retweeters of a politician and calculate a score for each topic (aka hashtag) he or she is talking about. This helps to identify, if a tweet is really interesting to the public or if it is only interesting for certain bots.
The list of politicians was loaded from github.com/okfde/wahldaten.
pip install pipenv
pipenv install
cp .env.sample .env
# now edit the .env to match a ArangoDB installation (or contact
# @codeforfrankfurt on GitHub or @codeforffm on Twitter for the one we use).
# If you want to query twitter data, you'll need to specify credentials for a created
# Twitter app as well. Then you can run twitter jobs that save to the DB you configured.
cd polbotcheck
python twitter_api.py --all
# or
python db.py --all
# To run the API for the frontend
cd webapi
cat README.md
cd web
npm install
npm start
npm install is only needed the first time.
-
Arango Server Commands:
arangodump --server.endpoint tcp://disruptivepulse.com:6754 --server.username root --server.database polBotCheck --output-directory "dump" arangoimp --file SOMEFILE.json --server.endpoint tcp://disruptivepulse.com:6754 --server.username root --server.database polBotCheck --collection users --create-collection true
Follower: twitter follower concept
Engaged user: User B is engaged with user A if user B retweets or likes user A tweets.
Audience: the ensemble of the set of followers with the set of engaged users in a specific topic or domain.
Bot score: belief on how much a user is a bot.
Botness: or bot level is a measure of the weighted ratio of believed bots in a given entity (e.g. user's set of followers, or user's audience).