I used the attached data source of shutter access logs.
shutterbot log.xlsx (214.6 KB)
I cleaned up irrelevant data.
I excluded logs which showed only an open without a corresponding close.
The infographic is entirely ChatGPT generated.
I haven’t validated the output, so I would be very happy if someone validated a few of the data points.
I would also like some comments on its utility - I mainly created it to address open evening questions about how easy it is for non-shutter holders to open the space.
3 Likes
There’s no way i can find the time right now to validate that data, so i too used ai; GLM-5.3-Flash. I felt that perhaps, if I approached it the way I would typically if I had done this by hand, then there’s a potential for a useful comparison between what ai generated for me verses what it generated for you. However, at the last step I was hand wavy and just told it to calculate all fields, so I didnt really hand hold besides when I noticed it didn’t look right to me and so corrected it a bit.
I used one session to clone your infographic and populate it with sample data to have a foundation to work with. Also generate a markdown file explaining how the data should be populated by the next step.
slms-infographic(1).zip (256.9 KB)
Then took your csv and had it whittled down to something i felt more usable. Each row being a single day, with events for up and down shutter control in the same row - and hours open counter. i tried to encourage it to be aware of the shutter opening/shutting on subsequent days.
shutter_daily_open_hours_v2.csv (159.8 KB)
Then directed another session to take the input csv, disregard the period of no shutter access logged as that was skewing previous results, and to insert the result into the infographic html.
slms-infographic-real(1).zip (267.4 KB)
To iterate on/reproduce.
shutterbot-repro-pack.zip (1.4 MB)
Thanks for doing this. Not a million miles between them! Makes me a bit more confident of the results.
1 Like
