[06:15:45] [telegram] Hello everyone, I have a good nwes for all .If you are interested in crypto currencies trading like USDT contant me, I can explain how can you trade and we can work toghther. If you would like to sell USDT，I can pay higher amount than normal 。Don't miss the chance guys.I am waiting for you .Have a great day. [09:37:48] [telegram] Is there anyone here who could do something about T263908? I just noticed the article is given a quite prominent spot on the front page when you open the Wikipedia Android app, which might partly explain why someone is going to such lengths to inflate its pageviews [14:25:29] [telegram] Maybe add it to blacklisted entries in the app? (may require creating a blacklist if one is not existing so far) (re @jhsoby: Is there anyone here who could do something about T263908? I just noticed the article is given a quite prominent spot on the fro...) [14:31:26] [telegram] Well some things end up being inflated. For example on Ukrainian Wikipedia many of most popular articles are Ukraian and Russian state media: the views that clearly go from YouTube's warning that the channel is sponsored by a state, and thus we usually exclude those entries along with other suspicious ones when providing most vied lists to the media and such [14:38:19] [telegram] Well, that case is still different from Puigdemont since it is real people coming there. As can be seen on that page statistics there is not a natural distribution from mobile and desktop views but clearly a tool hammering the servers. (re @Thecladis: Well some things end up being inflated. For example on Ukrainian Wikipedia many of most popular articles are Ukraian and Russian...) [14:40:01] [telegram] I wonder how this didn't get solved and that flower image mystery did [14:40:39] [telegram] (meaning T273741) [14:41:13] [telegram] not on the right radars i guess (re @mahir256: I wonder how this didn't get solved and that flower image mystery did) [15:07:16] [telegram] Thank you for mentioning this ticket. I had a good laugh while reading it 😄 (re @wikilinksbot: T273741 – Investigate unusual media traffic pattern for AsterNovi-belgii-flower-1mb.jpg on Commons [resolved]) [15:21:30] [telegram] Scale, that flower was used a LOT. 90 million times a day (re @mahir256: I wonder how this didn't get solved and that flower image mystery did) [15:24:55] [telegram] Its rather ineffective as marketting strategy as well. Last I checked the traffic coming from the apps is small. (re @mahir256: I wonder how this didn't get solved and that flower image mystery did) [15:25:36] [telegram] So is @jhsoby's renewed concern overblown then, in the grand scheme of things? [15:28:17] [telegram] I guess the real bug is the apps prominently showing top viewed articles in the first place? [15:32:09] [telegram] I think the problem is that this top 3? 5? on smaller wikis is just very inaccurate and very sensitive to any bot engagement/errors. [15:32:32] [telegram] as such, I think that on smaller wiki's, it's probably better to make a random selection from the top 20 or something. [15:36:05] [telegram] It's boring, is there anyone to chat 😀 [15:36:28] [telegram] This isn't just smaller wikis, the article is consistently in the top 10 for many large European languages (re @djhartman: I think the problem is that this top 3? 5? on smaller wikis is just very inaccurate and very sensitive to any bot engagement/err...) [15:36:41] [telegram] Its gamed in ptwiki as well (re @djhartman: I think the problem is that this top 3? 5? on smaller wikis is just very inaccurate and very sensitive to any bot engagement/err...) [15:38:42] [telegram] It looks like someone (Puigdemonts PR-manager?) has found a very effective SEO tool and is abusing it. It would be nice if we could put a stop to it. [15:39:07] [telegram] Is this effective SEO? (re @Tohaomg: It looks like someone (Puigdemonts PR-manager?) has found a very effective SEO tool and is abusing it. It would be nice if we co...) [15:39:12] [telegram] sounds like something the Spanish government might say /s [15:40:14] [telegram] For ptwiki the big one using this is AMBEV, seems Sony Channel recently joined in. [15:41:13] [telegram] Yes. To make sure that nobody forgets about the "opposition leader persecuted by the Spanish government". (re @chicocvenancio: Is this effective SEO?) [15:41:39] [telegram] I'm sure that is the point. Not at all clear this is effective. [15:43:01] [telegram] https://tools-static.wmflabs.org/bridgebot/b8b4e0fb/file_11188.jpg [15:44:46] [telegram] The little green bar is mobile app views. Still ~200 -800k views per month, but far from effective SEO. And that is the potentially impacted views, surely only a small fraction of it is influenced by the "Most read" pannel in the apps. [15:47:05] [telegram] but which search engine does it optimize for? (re @Tohaomg: Yes. To make sure that nobody forgets about the "opposition leader persecuted by the Spanish government".) [15:49:01] [telegram] And that is across all projects, several wikis have a much smaller fraction of app traffic. Basque doesn't even show up in the bar graph. [15:52:52] [telegram] circling back to this. As a bug for the app, it is surely very relevant and I think its another facet of our terrible support for mobile as whole and apps specifically. As an issue for the movement itself, yes, not a big deal. (re @mahir256: So is @jhsoby's renewed concern overblown then, in the grand scheme of things?) [15:56:10] [telegram] I've tagged the ticket with the analytics project and their kanban board [15:56:32] [telegram] that at least gives it visibility among the correct people (page views is just a subset view on top of analytics) [15:59:06] [telegram] What is funny, is that you can see a dip in September, which I suspect is the lets encrypt issue, which broke their tool. ;) [17:00:51] [telegram] Hi [20:36:00] [telegram] Has anyone here played with supervised NER matching using a subgraph of Wikidata?