The program takes an html file of the youtube history data and converts it into a Pandas dataframe that can be exported.
The html file can be obtained from a Google user's Takeout page by following the instructions in this video
Obtaining an API key from Google Developers adds the length of the videos watched to the dataframe.
One of the challenges encountered in this project was using regular expressions to obtain video urls, watch dates, times, and so on.
I would love to do some analysis on the patterns in my own data. For example, what channels I watch the most in a month, how much the time I spend on Youtube changes over time, what genre of videos I like, and so on. I would have loved to have information about repeated watchings of the same video, but the history data resets every time a video is watch.
- Adding OAuth so that a person can just run the program and sign in to their Google Account
- Adding a Bokeh plot to generate dynamic user-determined plots. For example, the most viewed channels in a specified time interval.
- Generating a word cloud of most common terms in video titles