>>8728878every site on the internet knows your user agent. your browser voluntarily presents it to the site when making an http request.
anything you connect to knows your ip address because it needs to know your ip address to actually make a connection. if it didn't know your ip address, it wouldn't know where to actually send the data. this is like one of the most fundamental parts of how the internet works.
as far as geolocation, that's usually an approximation based on where your ip is located. it's usually pretty inaccurate, it's going on the known location of one of your internet service provider's endpoint nodes, which if you're lucky is close by.
if you're on a cell phone it might request the coarse or fine location permission (which you have to approve). coarse uses triangulation based on what three cell towers (from any service provider, not just the one you actually use) are closest to you, which usually gets your location reasonably accurate (incidentally in the us this is the only location data ems has access to, they can't use your gps when you call 911, it's a major problem). fine uses gps, which is accurate to within about 100 feet usually.
as far as the first statement, it sounds like they store search terms without a unique identifier. that makes "we don't collect personal information" a bit sketchy. if you consider search terms personal information, then they are collecting it, just without a unique identifier. however, there are other ways to actually track this, like the three pieces of data at the end. if you have someone's user agent, ip, and general location, you can usually uniquely identify them. and since this information is made available to just about anything you connect to, there's no guarantee that they're not tracking you other than their word.
so i'd say be skeptical. but the infographic on its own doesn't really prove anything. you're never going to be guaranteed of safety unless you run your own webcrawler and search engine.
