I noticed something very odd about Linkedin anti-bot behavior and I am not being able to fully understand it. Basically, there are multiple people who set their profiles to public, allowing search engines such as Google to index them. If you initiate an anonymous browsing session and search for a profile, you will get a page that looks like this:
(example for https://www.linkedin.com/in/parag-agrawal-5a14742a)
If you try repeatedly, however, or if you test an automation software such as selenium for accessing these profiles, you will after a while be redirected to an authwall page, and no more public profiles will be visible to you.
I have already tried to hide all kinds of fingerprintable data, such as IP, user agent, screen resolution (for canvas fingerprinting), but after a few requests, my access limit drops to one single public profile. I guess they make an analysis of the request pattern within their servers, making my section identifiable.
I would like to know which features are observed to determine whether a visitor is human or not, and how could I simulate human behavior.