Sybillian

What is it?

Sybillian is a twitter bot that can receive mentions or be included in conversations to analyze the account(s) to see what their bot score is based off of a BotOrNot score.

BotOrNot checks the activity of a Twitter account and gives it a score based on how likely the account is to be a bot. Higher scores are more bot-like.

BotOrNot accomplishes this goal by the use of Random Forest to do ensemble learning via a multitude of Decision Trees. The model is then trained with 'known bots' and 'known human' accounts to build a heuristic decision tree model that nullifies bias. Sybillian's role is to present BotOrNot accounts to test against this ensemble and provides a score for multiple factors, including:

• content classification
• friend classification
• network classification
• sentiment classification
• temporal classification
• user classification
• language agnostic classification

For further details, you can reference the Indiana University Publications.

What is it's purpose?

I built Sybillian because I wanted to offhand check the "truthiness" of accounts that were pushing a specific topic or idea. Whether it be for marketing or otherwise.

How can it be used?

You can either directly mentions @SybilDetector with a username. ex:

@SybilDetector @SwiftOnSecurity

and receive a response back:

@Decad3nce SwiftOnSecurity has a bot score of 0.45. Classification of 'might be a bot'. See
https://t.co/upkHRBwQ4S

Or, you can add it as a reply to a twitter "thread" to see the percentage of participants (including retweeters) ex resp:

@Decad3nce the conversation of 76 contributors contains 4 probable bots, percentage of bots in conversation 5.26315789474.

How can I deploy my own?

1. Create a Twitter Application.
2. Populate credentils.py with your tokens and shared secrets from the application creation.
3. Run locally by creating a virtual environment and pip install the active directory.
4. Run remotely by hosting the application on whatever server of your choosing.

Isn't this just a shell for BotOrNot?

Yes, yes it is. But it also allows the process of 'harvesting' bot data to be further decentralized.

Future plans?

1. Move away from local sql db to a generic and pluggable backend for caching that can be hosted remotely. Then allow surrogate bots to provide data to that central cache.