We collect only two pieces of data from each inbound request:
(1) the time the request was made
(2) the path of the request, for example
The vast majority of mainstream analytics platforms pay little regard to privacy. Their policy is to collect as much data as possible, satisfying the site owners' desire for information to the detriment of the end users' privacy.
So, we built a simple Rack application that collects only the data that we want to collect: the number of requests hitting each page on our site, and no more. It's open source! See how your data is being collected.
*Except where unavoidable for the purpose of web server logs