workspace-analysis

Helpful Engineering Slack data analysis toolkit

A comprehensive set of tools for analyzing the contributions to the Helpful Engineering Slack community. Currently it's focused on model generation for a bot-based channel recommendation system, but is versatile enough to generate even channel-specific word clouds, per-channel activity charts and some other niceties.

📖 Usage

1. Install the requirements:

   pip install poetry
   poetry install

2. Run the examples:

   poetry run python -m examples.channel_list --no-cache --token xoxp-··· \
                                   --format csv --output ./channels.csv
   
   poetry run python -m examples.classifier_model --channel-filter "(project.*|skill.*|communication.*|discussion.*|hardware.*|medical.*|legal.*|comms.*|fundraising.*)" \
                                       --channel-threshold 0.5 \
                                       --output ./model.json
   
   poetry run python -m examples.data_visualization --output ./images_folder

Note: the first run may take eons while gathering the information.

🔒 Security

After the first run, ./corpus/cache is populated with sensitive data, and it should be handled with the same care as the token.

---

Future improvements

• Integrate tool-experience-tagger into this project.
• Automate the category training method or, at least, make it easier through prodigy or doccano.