🔍
RnDAO workspace
/
🗣️
WeLivedIt.AI
/
🗓️
WLI weekly D+Z
🗓️

WLI weekly D+Z

Status
Date
Apr 15, 2025
summary
Discussion on steps needed to getting some results for the “Discord to Score” project that can be presented on Friday.
Todos:
Provide simulated discord data Who: Adedamola Deadline: Wednesday afternoon Notes: short sample to be ready earlier
Python script that processes the data and gives back some statistics / descriptions of hs/toxicity that can be displayed on the dashboard. Who: Zaremba Deadline: Thursday morning Notes: provide partial update to Adedamola by Wednesday afternoon
 
More details of the meeting:
First Zaremba apologised for not providing promised materials before the weekend, as they lost quite some time being ill.
There was a short discussion of a trained community model as a data product. It was noted that it is useful to have this in mind, but not as a new focus, but as an extra narrative around our primary focus of providing tools for community moderation. 
The main discussion was about steps needed to be able to offer a presentation by Friday (morning?).
Adedamola has already provided Zaremba with a small dataset from WLI discord that contains the hs/toxicity testing channel. Zaremba is using this as a starting point data. Adedamola will provide Zaremba with a simulated discord data. And Zaremba will then provide Adedamola with a python script that processes the data and provides output in the form of statistics / descriptions to be incorporated and shown on the front end (dashboard).
The simulated data will: • Have the same format as the small WLI discord data (comma delimited csv file, with the following column titles: Guild, Channel, Author, Content, Timestamp) • Goal for the size: several thousands data points (messages), with several percents of those containing hs/toxicity • Community: web3 hackathon community/chat, safeguarding focus: sexism and racism • Channels should include: team formation, workshops, general chat, technical questions • hs/toxicity examples should include personal attacks towards a female hacker wanting to join a team (“We’re looking for programmers, not models”, “Make me a sandwich”), and towards a workshop instructor who is a person of colour (“That instructor was surely a diversity hire!”)
🔍
RnDAO workspace
/
🗣️
WeLivedIt.AI
/
🗓️
WLI weekly D+Z
🗓️

WLI weekly D+Z

Status
Date
Apr 15, 2025
summary
Discussion on steps needed to getting some results for the “Discord to Score” project that can be presented on Friday.
Todos:
Provide simulated discord data Who: Adedamola Deadline: Wednesday afternoon Notes: short sample to be ready earlier
Python script that processes the data and gives back some statistics / descriptions of hs/toxicity that can be displayed on the dashboard. Who: Zaremba Deadline: Thursday morning Notes: provide partial update to Adedamola by Wednesday afternoon
 
More details of the meeting:
First Zaremba apologised for not providing promised materials before the weekend, as they lost quite some time being ill.
There was a short discussion of a trained community model as a data product. It was noted that it is useful to have this in mind, but not as a new focus, but as an extra narrative around our primary focus of providing tools for community moderation. 
The main discussion was about steps needed to be able to offer a presentation by Friday (morning?).
Adedamola has already provided Zaremba with a small dataset from WLI discord that contains the hs/toxicity testing channel. Zaremba is using this as a starting point data. Adedamola will provide Zaremba with a simulated discord data. And Zaremba will then provide Adedamola with a python script that processes the data and provides output in the form of statistics / descriptions to be incorporated and shown on the front end (dashboard).
The simulated data will: • Have the same format as the small WLI discord data (comma delimited csv file, with the following column titles: Guild, Channel, Author, Content, Timestamp) • Goal for the size: several thousands data points (messages), with several percents of those containing hs/toxicity • Community: web3 hackathon community/chat, safeguarding focus: sexism and racism • Channels should include: team formation, workshops, general chat, technical questions • hs/toxicity examples should include personal attacks towards a female hacker wanting to join a team (“We’re looking for programmers, not models”, “Make me a sandwich”), and towards a workshop instructor who is a person of colour (“That instructor was surely a diversity hire!”)