Yeah, sample bias is always a risk, just like it was with the GameStop thing. But I think it slightly misses the point. Even though we think of social media as something we voluntarily participate in, it goes way beyond that, from CCD cameras all over the city to popularity-based recommendation algorithms, face recognition scraping, Google maps location tracking, etc.
E.g. The Jester used to have a sentiment analysis tool that examined people's online writings, tweets, etc. that assessed whether or not that person had violent tendencies. And, as I commented before, I'd like to know whether Tom Trenchard is an alias for S. Adam Seagrave, which I think could be determined with techniques similar to those used here. On 2/9/21 12:21 PM, Marcus Daniels wrote: > Possibly an interesting inference for the subset of people that are > transparently social on social media. As opposed to the people that don't > use social media, or only use it reluctantly for professional things, or use > it to exaggerate or pose themselves or their family. These other groups > will have little or zero revealed. Annoyingly, the measured will tend to be > converted into the normal. -- ↙↙↙ uǝlƃ - .... . -..-. . -. -.. -..-. .. ... -..-. .... . .-. . FRIAM Applied Complexity Group listserv Zoom Fridays 9:30a-12p Mtn GMT-6 bit.ly/virtualfriam un/subscribe http://redfish.com/mailman/listinfo/friam_redfish.com FRIAM-COMIC http://friam-comic.blogspot.com/ archives: http://friam.471366.n2.nabble.com/
