[piqi]

Your site works great when permitting only first-party images/css/html. I use Firefox's tracking protection + ublock origin + umatrix. I don't allow first-party JS by default, or any third-party stuff. I spoof the referer. I see JS from plausible.io tries to load.

I wonder how that skewed the requests. Do you have data on that? I'm sure there are some % of HN visitors that do the same. Less likely the case for Elon Musk followers. I don't know how you would get 'time on page' data without JS.

[simonw]

The best data I have on that is the delta between Cloudflare numbers and Plausible numbers - but that's not quite the right comparison, because for Cloudflare I have all traffic to my site, not just traffic to that specific page (which I can get from Plausible).

[piqi]

What about device OS or browser data? User-agent should be enough for a meaningful breakdown.

People mainly use the twitter app. HN is just a website. I'd expect more desktop users from HN.

EDIT: oh I see. It's in the github repo and vercel app.

name visitors

Mobile 778,678

Desktop 101,216

Laptop 47,781

Tablet 16,967

[simonw]

Yeah the device data from Plausible is here: https://i-will-not-harm-you-unless-you-harm-me-first.vercel....

[pclmulqdq]

I think you can reasonably assume that the cloudflare traffic from that time is almost entirely headed to this page, minus your baseline number from other posts. The "total requests" will be completely skewed, but "unique visitors" should be okay.

A lot of the HN crowd seems to use adblockers and script blocking.