zlacker

Screen sizes is ambiguous here - are you measuring viewport width (`window.innerWidth` - helpful) or the display the window happens to be on (not too helpful)? Also something to make that data useful would be show the range of sizes, instead of the top specific size. E.g. 1280 may be _the most popular_ but there may be more users using larger width windows, just more variation in those sizes (1320, 1440, etc), so a top level range could be a nice differentiator here.

But, how useful are these stats going to be without being able to see user journeys through a path of pages / actions? Yes, it's good to know which pages are getting how many views. But, in order to improve the UX, we often need to know how many users are able to go from Page A to Page C and whether they went through Page B first. Or e.g. if 90% of sessions that start on Page A (so we know what their purpose was), end on Page B but the main (perhaps beneficial) action for the user was on Page C. You can't just look at the pageviews for each, because you don't know where the session started.

I fear that this would reduce people to "inferring" (guessing) too much about the data that they see, and making decisions they feel are backed with data when there's not enough data to conclude. Then again, I'm sure that happens when the data is there too :-)