Heat Detector

Heat detector is a StackApp whose goal is to catch offensive, rude, snarky comments, and comments that leads to abusive content (offensive, spam posts, or previous offensive comments). It is running a bot that queries comments from the Stack Exchange API and retrieves all comments posted on Stack Overflow.

The automatic detection is based on a mix of regular expressions and NLP (Natural Language Processing). When a comment is classified as “rude” by the system, it is sent in a dedicated chat room for manual review and feedback.

Example

output

To avoid storing comments in chat transcript (unearth buried fights), the comment is automatically deleted from chat before the timeout (which is a 2 minute window).

The bot also has a test function and a report function, that enables the possibility to test different comments and to train the classifiers on uncaught content.

Guidelines

Stack Exchange has provided a feed of offensive comments which allows for testing and serves a base for the classifying algorithms. However, do not use this as a witch hunt to find users and harass them.

  1. Commenting back on a SO post to a heated comment is, in general, a very bad idea. Avoid commenting in most, if not all cases. A Room Owner might decide to stop the bot if the room is involved in any bad situation (using the command @Queen stop feed) and kick the users involved.
  2. It is not allowed to speak about the users that posted reported comments in the chat room. We do not moderate users, but comments.
  3. When multiple offensive comments are exchanged or someone is going well over the top, please remember to flag the thread for moderator intervention. They are not automatically notified of the comments’ deletion.
  4. tp should be used for “not nice” comments or comments that lead to offensive ones or offensive / spam posts (answer / question). You can flag the comment or the post approprietly.
  5. tp socvr can be used to send a notification to the SOCVR room. This should only be used when the comment is very rude and requires immediate attention to avoid flame posts and / or repeatingly offense from other users.
  6. fp should be used for comments that were wrongly reported (i.e. they are on-topic and constructive).
  7. It is up to you to give feedback when and if you want to. You can abstain from giving one if you feel the comment is hard to classify (borderline).
  8. As a general guideline, please don’t pile on flags unless the comment really warrants it. Only one flag is enough for a moderator to take a look and act on it.