Chronotope’s avatarChronotope’s Twitter Archive—№ 153,498

  1. …in reply to @burk504
    burk504 It's just the Topics that Google's classifier have associated with the domains you have seen the user on. If you're running the script on a single domain you're just seeing at least one Topic the classifier has determined is related to the domain and likely at least one random...
    1. …in reply to @Chronotope
      burk504 Considering that for the duration of the test you can see the classifier's actual designated Topics per domain, I'd bet that everything but the top one, maybe two, topics on your list are randoms. Here's the actual classifications of domains from topic-internals...:
      oh my god twitter doesn’t include alt text from images in their API
      1. …in reply to @Chronotope
        burk504 And for comparison if I visited your domain you'd likely get the actual topic plus one or two of the random Topics the API has for me over the last 3 epochs -
        oh my god twitter doesn’t include alt text from images in their API
        1. …in reply to @Chronotope
          burk504 I am somewhat baffled that my domain that stores archived copies of articles I read (which are almost never about Vehicles, Vehicle shopping or shopping) is classified as 83. and 289. though.
          1. …in reply to @Chronotope
            burk504 The other thing that can happen is that you can visit a domain and not get a Topic calculated at all, from my testing. But I'm not sure what condition causes that.
            1. …in reply to @Chronotope
              burk504 The user only gets 5 topics per epoch and I guess visiting a page once or twice may not be a strong enough signal to attach a topic to my user agent.


Search tweets' text