An application to analyze Zoom meeting text
Did you know that when a Zoom meeting is recorded the chat is saved as well (including the private messages)? I recently called in to the annual meeting of a research organization I was a part of during my master’s. In parallel to the great presentations we had an excellent discussion going on in the chat box.
When I think of a Zoom recording I normally only consider the video recording but all sorts of information is contained in the chat transcript as well! New project ideas, interesting web pages, insightful questions. Not to mention the insights about the speakers or participant community we might be able to find. I figured someone must have created a tool for Zoom chat analysis for just this purpose but if they have, I haven’t found it. So, using my newly crafted data science skills, I decided to make my own.
The new text analysis tool currently runs as a terminal application that can:
1.Create and export a csv from the ZoomChat.txt file with three columns: time of comment, author, and comment body (excluding private messages)
Know some NLP and analytics? I’d love to see what you can build with the organized text data! I also have a whole pile of ideas for deeper analysis but in the meantime I started with a few basics.
2. List which participants commented the most
3. Create and export a network visualization
While knowing who chatted the most is good data, seeing the nodes of who talked or responded most to whom can provide further insights into the community structure of participants.
4. Create and export a histogram of comment volume over the course of the meeting
When were the most comments sparked? The least? If the meeting was a scheduled event we can compare the histogram with the event timeline to answer questions like which speakers or subjects caused the most discussion. Here, for example, we can clearly see the mid meeting break at 10:20.
5. Extract any URLs
Don’t want to lose track of all the amazing resources participants shared!
6. Export the csv, run functions 2–5, and save the output from functions 2–5 into one pdf report.
If you are the administrator of a business Zoom account you can see some of theses analytics in the Zoom Dashboard. But this application is free, does more analysis within the text than the Dashboard, and works for those of us without memberships.
To try the application , see more example output, probe the inner workings of the code, or collaborate, please head over to my GitHub repository: https://github.com/LVParkinson/Zoom_chat_analytics
What’s next? I’m currently working on adding a function for sentiment analysis but I may have to find a different transcript to play around with. The annual meeting was a pretty positive event. I need a more diverse sample!
Special thanks to the Bonanza Creek Long Term Ecological Research Program for the inspiration and access to transcript text.
