What is Chat Filter?

Chat Filter uses Slack’s API and Microsoft Azure to detect when users post negative or non-inclusive messages to Slack, in which case the channel is notified that the message could be more appropriately worded.

This project can be customized to allow more discrete reporting to supervisors and fine-tuned for any team.

This was built during Local Hack Day 2018 at the University of British Columbia.

We’re Proud of:

We entered Local Hack Day with no JavaScript experience, yet we’ve used it to create something awesome.

We were really excited to learn about Microsoft Azure, and we eventually settled on and successfully used its Text Analysis API.

Many of us had previously found Slack to be a helpful tool, so we are excited to have learnt more about it from a developer point of view and proud that we created something that is useful to us when we use Slack in the future.

We left Local Hack Day as a team of 4 friends, and we think Chat Filter has some awesome use cases.

What challenges did we face, and how did we overcome them?

Our unfamiliarity with JavaScript blocked us for awhile because some of its nuances stopped our HTTP posts from working.

Nevertheless, with three of our mentors, Conrad Muan, Albert Xing, and Alicja Raszkowska provided us with appropriate alternatives and solutions so that we could continue.

In addition, we learned to parse JSON objects that were required for interaction with Azure, for which we want to give a shout out to well-done documentation @Microsoft.

Future Ideas:

Chat Filter could reply in threads instead of the channel directly in order to minimize visual clutter.

Customize Positive and Negative Messages

Team:

Erik Ryhorchuk

Leo Zhang

Moses Kirathe

Melissa Bernstein

Resources

Harassment Detection

Harassment-Corpus

Share this project:
×

Updates