How can I see statistics about when the number of views increased over time?
I would like to find out if at a certain date there was a significant increase of views to one of my "famous" questions.
Add a comment
|
2 Answers
This is not possible without the use of external tools. Historical view counts are simply not stored. One such tool would be the Wayback Machine; you'll see some of the more famous questions like this one of yours are already indexed now and then.
| Date | Views |
|---|---|
| 2021-01-24 | 222k |
| 2021-03-05 | 225k |
| 2021-06-20 | 233k |
| 2021-07-07 | 234k |
It's relatively easy to make a curl script (it's just a GET call to https://web.archive.org/save/[insert URL here]) to save a page to the Wayback Machine, and you can schedule that as a cronjob, as to have a more complete coverage.
@Trilarion suggested using votes, which is indeed the most accessible proxy for which we can construct historical patterns through the Stack Exchange Data Explorer. If you're happy assuming the number of views are proportional to the number of votes (up and down), this query gives you the result you want. As I noted in the query description, I don't think this is a reliable method, but some people may like it.
answered Jul 12, 2021 at 6:24
-
The historical view counts aren't stored statement is from 2015. That could have changed by now. It would surely be meaningful information. As another proxy one could maybe take votes as proportional to a certain number of visitors and for votes there would be timestamps. Commented Jul 12, 2021 at 21:24
-
4@Trilarion It has not changed and is unlikely to ever change. Commented Jul 14, 2021 at 16:53
-
@animuson Thanks for the confirmation. That's a quotable recent source. As for the second part, let me just say that I don't understand it. Of course you are free to do whatever you like, but I think it's actually useful information. Commented Jul 14, 2021 at 18:01
-
1@animuson FWIW system could sample view count on some reasonably low frequency. Once a day would suffice for the purposes of historical analysis. Maybe even once a week would do the job– gnatCommented Jul 14, 2021 at 19:17
-
3@gnat I'm actually doing something like that for my own questions.– Glorfindel ModCommented Jul 14, 2021 at 19:31
-
Rather than using votes, using anonymous feedback would be more representative of the general population, given that 99% of the hits are by anonymous users.– BraiamCommented Jul 16, 2021 at 15:35
-
1@Braiam that's a good idea, but I was surprised there is less feedback than actual votes; not only on this post, but also in general: data.stackexchange.com/stackoverflow/query/1436074/…– Glorfindel ModCommented Jul 16, 2021 at 15:51
-
1Of course it would be less, because you are counting all votes not just up/downvotes. They are less but not by several orders of magnitude. data.stackexchange.com/stackoverflow/query/1437168/…– BraiamCommented Jul 19, 2021 at 15:51
-
@Braiam good catch! Still, instead of 4x the amount of feedback, it's 3x the amount of feedback. With so many anonymous visitors (not sure if it's 100 to 1, but it's certainly 10 to 1) I'd expect more feedback than votes.– Glorfindel ModCommented Jul 19, 2021 at 17:53
-
Yeah, I'm trying to figure out why the difference, because I usually see posts with tons of user votes, with many more anonymous votes.– BraiamCommented Jul 19, 2021 at 21:21
As of February 2013, this is technically not possible since views aren't tracked over time. (source: https://meta.stackexchange.com/a/168230/169611)
However, if the question and its answers have been accumulating a lot of votes, you can look at the timeline of the question to see if there any spikes.
An example timeline is: https://stackoverflow.com/posts/5859561/timeline
In this example, you can see that the question got a lot of attention on May 4,2012
In your case, it's likely that your two Famous Questions are attractive search engine hits. So people are coming via Google and such. In this case, the views would be coming in steadily over time and not in a single spike.
answered Apr 29, 2014 at 18:33
-
2