-2

I am trying to collect some data for a pet study. I would be collecting some metadata on the video suggestions based on a search. I was wondering if it is possible to do the following using the Youtube API(python or R) :

  1. Input a search keyword and get the results
  2. Choose one of the videos randomly and see the list of video suggestions.
  3. Choose one of the suggested videos randomly.
  4. Repeat this "n" times.

Is there a way to emulate this entire process? I think web-scraping can be an option but I am not really sure, how I would go about it. So if there are nay pointers that would be amazing and get me started.

Also, Is it possible to have no history, as in an option to erase all the cookies for the previous attempt(Steps 1 through 4) and start afresh? (More like an option to have this in Incognito Mode)

TIA for your suggestions.

Ishan mistry
  • 1
  • 2

1 Answers1

1

AFAIK Google tracks the computer you are using in such a way that you can't escape their filter bubble. Even through Tor, YouTube might prefer some content related to the exit node IP location (and so language) or any previous YouTube search done by you (through this exit node) or another user of the end node or any computer using the same IP as the exit node...

The YouTube Data API v3 has a possibility to retrieve suggestions thanks to part=suggestions with Videos: list by authenticating with OAuth (so results might not be neutral). You can get the initial videos thanks to Search: list thanks to q filter. Web-scraping is also doable to be less tracked, my open-source YouTube operational API is able to web-scrape search results for instance.

Note that a French person claims having achieved to have done such a neutral French YouTube suggestions graph.

Benjamin Loison
  • 3,782
  • 4
  • 16
  • 33