Talk:Hamme (disambiguation)

on incoming links
I've gone through most of the incoming links and found ambiguous links to the river, and numerous references in context to East Flanders (VOV) or Flanders. I've used topical redirects to fix these, and most of the references were disambiguated in text already. I will now disambiguate the term fully because it's unlikely the average English reader associates the term with the town as opposed to the river, or other largely homonymous terms. --Joy (talk) 09:04, 20 February 2024 (UTC)

post-move
In December '23, we saw 306 incoming views at Hamme, and in January '24 a spike to 572.|Hamme_(disambiguation)|Hamme,_Belgium The clickstreams do not register anything for those months, though, so the hatnote couldn't have gotten more than e.g. 10/traffic or e.g. 4*9/traffic clicks, so ~3-12% for December or ~2-7% for January.

In March '24, we saw a total of 176 views at Hamme, and in turn:
 * clickstream-enwiki-2024-03.tsv:
 * Hamme       Hamme,_Belgium  link    28 (~16%)
 * Hamme       Hamme_(river)   link    10 (~5.6%)
 * total: 38

All of this is too close to the anonymization thresholds to be able to make strict comparisons, but it doesn't seem like the Belgian town was really the primary topic (because it's hard to assume e.g. 28 + 9 + 9 + 9 = 55 and that's a far cry from 176 anyway). --Joy (talk) 09:53, 7 April 2024 (UTC)

If "Hamme" gets 176 views in a month, and "Hamme, Belgium" gets 246 views over the last 30 days, then how was Hamme, Belgium not the primary topic? The average pageviews dropped from 16 or 18 to 6 here, while the average pageviews for the ",Belgium" page are at 8. Average pageviews for Hamme (River) are 1, for the football team 3, and for Hamme, Bochum 0. It seems very obvious that Hamme, Belgium is and was the primary topic here. Fram (talk) 08:52, 9 April 2024 (UTC)


 * This is the same logic as I just explained at Talk:Rasna, but in this case it's even more pronounced because the observable difference between the town and the river is even smaller and we're even further away from the entirety of the readers. --Joy (talk) 10:08, 10 April 2024 (UTC)
 * To your first question - raw page views do not translate perfectly to average reader behavior. A lot of our traffic comes from external sources, such as search engines. Many readers will have a session with their search engine which will indicate that e.g. they're from Belgium or that they're taking an interest in topics related to Germany, and the search engine will most likely do its best to short-circuit them according to their location, or in turn other characteristics, or in turn maybe it will do something based on some other interests (maybe a reader who reads about a place is later more likely to look up shops in that area, leading to possibly more monetizable traffic with the search engine?). That's why I find the clickstreams slightly more indicative than view statistics, because they hint to us the intents of the readers as they are presented with information without an obvious third-party to influence traffic. --Joy (talk) 10:15, 10 April 2024 (UTC)
 * The views without clickthrough might just as well come from webcrawlers, indicating nothing about our readers as well. The only thing we are certain about is that the vast majority of readers who do click through are interested in the town, and that this matches the relative pageviews of the target pages. Which is what WP:PT1 indicates as a good primary topic. Fram (talk) 10:41, 10 April 2024 (UTC)
 * No, I'm afraid that is just not correct, because meta:Research:Wikipedia clickstream documents:
 * We attempt to exclude spider traffic by classifying user agents with the ua-parser library and a few additional Wikipedia specific filters.
 * We are not certain about that, primarily because we're looking at unreliable numbers, as I mentioned before. Again, from the above:
 * any `(referrer, resource)` pair with 10 or fewer observations was removed from the dataset.
 * That means any such outgoing clickstreams are missing, so when all the numbers we're looking at are so close to 10, we're just not looking at the complete picture and can't be certain about it.
 * And in turn, even those numbers that we see are just around 3 : 1. For an example of a recent discussion where this kind of a ratio was not deemed sufficient to move, please have a look at e.g. Talk:TOTP. --Joy (talk) 16:30, 10 April 2024 (UTC)