https://remoteindian.com/ logo
#random
Title
# random
c

chilly-lock-4986

07/10/2022, 10:10 AM
anyone who’ve exp working w data here? i got my youtube watch history in form of csv ( title, link, date ) and wanted to create something like spotify wrapped or similar visualizing this data
w

wonderful-dusk-20983

07/10/2022, 10:21 AM
What kind of insights do you want to track?
c

chilly-lock-4986

07/10/2022, 10:26 AM
most watched frequency ( figured this out ), trend of #videos watched over time, trends by artists, heatmap of day ( like at what time of day i was most active), most active day, least active day…
these are some of the things i was planning to track
w

wonderful-dusk-20983

07/10/2022, 10:50 AM
Not a specialist in data per say, but I think taking things from first principles - we need to add a few tags on the videos being watched For example Answering a question -
What was my most frequently viewed artist (assuming channel) on YT?
Tag here was channel being watched instead of video being watched We need to find the name of the channel for each video watched and cumulative sum of video frequency for a particular channel to come out to a ranking
If you have data like hourly usage or something, questions like heatmap of usage could be answered.
most active days, least active days would be a derivative of the above heatmap itself - add a tag called - total active hours per day
Hope this works, but if it does not do let mw know your confusions, I'll try resolving :)
c

chilly-lock-4986

07/10/2022, 12:55 PM
thanks, this was helpful, i figured out all the frequency ones. pandas really came in handy
figuring out plotting datetime related ones now, i think matplotlib should do it
also thinking of turning this into a side project web app soon
w

wonderful-dusk-20983

07/11/2022, 6:57 AM
Glad to know this helped
2 Views