Hi URA-ites! The bimonthly update will be on 4 February 2025 but I wanted to let anyone who's new or didn't know that the annual update of the historical data (here) will be updated by long time editor Altamel on 11 February. I think these data are really interesting and are great indicators of progress - you can see how many categories were cleared between 2021 and 2025, plus how ENORMOUS some of these categories used to be. I groan when I see a category with 500+ articles but in August 2009, almost every category was in the thousands. Really puts things in perspective and is a fun piece of Wikihistory. Cheers, Kazamzam (talk) 16:19, 29 January 2025 (UTC)
Headline: We cleared 2,963 articles and are now hovering barely over 69,000 (69,036 to be exact)! For yourself and your fellow editors, please clap.
Minutiae: For anyone interested in a more detailed breakdown of the numbers - average was 35.3 articles; median 12; mode 11. The average category declined by 6.6% and of the 189 categories being tracked, 149 (78.8%) had a decline of at least 10 articles and over 2% of their starting number compared to previous update (I consider this a growth metric that we're being comprehensive in clearing categories across the board). The smallest decrease was September 2019, which only decreased by 3 articles (I'm not mad, I'm disappointed). And the amount of time it takes to update all the categories individually is exactly the length of two (2) Simon & Garfunkel albums.
Highlights: Along with March and April 2009, December 2023 is in dustbin of history! This is the first non-chronological category clearance since 2022 (with the removal of July 2007). We also had a successful DYK submitted by @Cielquiparle for the show The Befrienders which had been unreferenced since 2006! And as editor in arms @Cakelot1 pointed out above, the unreferenced backlog is now under 1% of the total articles for the first time since April 2008. Huzzah.
Low-hanging fruit: We have a number of small fry thanks to the robust efforts of regular editors and drive participants. The infamous September 2019 is a humble 101 articles, dangling precariously like a sinner in the hands of an angry God. Give it a nudge.
High-hanging fruit: Everyone's favourite BFC (Big Friendly Category), December 2009, is a finger-lickin' 9,108 articles as of this writing, after a decrease of 460. The other high-hanging fruit are, still, the Frustrating Five (name open for revision): January 2013 (1,012), April 2019 (842), May 2019 (1,760), June 2019 (3,805), and September 2020 (1,004). This time, September 2020 had the lowest percentage of change between updates (1.57%). Godspeed to anyone working on these.
New challenge: No ties this time, but @ARandomName123 has set up a monthly leaderboard to encourage everyone's inner blood lust. Go forth.
Announcements: Per above discussion, the consensus seems to be that June 2025 will be the best time for our next drive. If anyone is inclined to start a draft, please be bold! Thanks, as always, for the amazing work. We're doing it! I'm calling it now - under 50,000 by the end of the year, if not sooner. All the best, Kazamzam (talk) 16:28, 4 February 2025 (UTC)
Hooray! Just wanted to quickly acknowledge everyone who contributed in January 2025 – every reliable citation is appreciated – and recognize the Top 10 editors adding references to completely unreferenced articles last month (more or less). Some impressive stats here:
Also - the next update will be in May 2025 so not to interfere with a) my vacation in the Japanese alps and b) the June drive. Hopefully we will knock through the better part of 2009, and then we can start to figure out how to scale Mount December 2009. Kazamzam (talk) 18:46, 4 February 2025 (UTC)
I'm aiming to kill off December 2024, I've been spending all my time working through there mostly alphabetically. So we'll hopefully have another non-chronological wipeout soon! SilverserenC23:15, 4 February 2025 (UTC)
Are these in-between drive month leaderboards still tracked using edit summary hashtags? Or are we able to get user names without that somehow? I'd like to participate! --Engineerchange (talk) 14:31, 16 February 2025 (UTC)
ARandomName123's awesome bot does it without your having to sign up. Just cite some unreferenced articles, remove the {{unreferenced}} template (maybe replacing it with a different one like {{refimprove }} if necessary) and that counts as one point to you. Cheers, SunloungerFrog (talk) 14:58, 16 February 2025 (UTC)
The bot counts it whenever an unreferenced article tag of any form (including norefs or nosources or any of the other variations used) is removed from an article. SilverserenC17:26, 16 February 2025 (UTC)
Just to clear things up, only the yearly one has the 5 article restriction. The monthly leaderboards start counting you when you do one. ARandomName123 (talk)Ping me!17:56, 23 February 2025 (UTC)
Quite humbling that there is a delta of ~200 between 1st place and 10th place. Mad respect to Silver seren in 1st, and well done to Kazamzam who managed to leapfrog Significa liberdade who was in 2nd for most of the month. Also impressive that Coldupnorth and Turtlecrown took 4th and 5th place, respectively, after taking a break in January.
To track where we are for the year to date, see the 2025 WikiProject Unreferenced articles Leaderboard, where you can sort by month. To join this friendly competition in March, add at least one reference to any article tagged as Unreferenced this month, and remember to remove the tag(s) after you are done; your stats will show up in the 2025 leaderboard once you've hit the 5 article mark.
Medicine-related articles about to hit a huge milestone
Could I beg for some help from this group? Wikipedia:WikiProject Medicine has been working for years to get at least one source into every article tagged by the group. We just have this short list left:
Most of these are organizations or other non-technical subjects. If you could please add sources to some and remove the {{unref}} tags, we would really (really really) appreciate it. WhatamIdoing (talk) 23:37, 6 March 2025 (UTC)
I wondered if there was any general good practice I should be following around sports databases (the likes of Cricinfo, Soccerway, Olympedia etc.) and where best to have them in articles about sportspeople. Up until now I have adding or moving them to the "External links" section, reserving the "References" section for inline citations from decent sources with significant coverage of the subject (e.g. news reporting and the like). But it occurred to me recently that they could also reasonably be placed in "General references", provided of course they're reliable sources, because they cover a lot of the tabular information that is often the majority of content in poorly-referenced sportspeople BLPs. I'd be grateful for any thoughts on this. Cheers, SunloungerFrog (talk) 14:41, 8 March 2025 (UTC)
I've been going through a lot of tennis tournament articles that are in the November 2024 category and I've been actively avoiding using any databases such as that. I don't personally consider them reliable sources and I do remember something about them not counting as proper coverage for sportpeople biographies, from the big re-organization of notability in that area a while back. So even if the databases are reliable, I'd rather be adding a source that does contribute something to notability for the subject. So I've largely been aiming for news coverage of the events in some fashion. SilverserenC17:08, 8 March 2025 (UTC)
Oh yeah, I always make sure I add one or more inline citations to a decent source. I personally don't count adding just a database source as proper referencing. It's more, if I'm there, and it's a football player, I might as well add in {{Soccerway}} somewhere if it's not already in the article, and whether that's better as a "General reference" or an "External link". (I use Soccerway as an example because an admin who does a lot of football content recommended it). Cheers, SunloungerFrog (talk) 17:39, 8 March 2025 (UTC)
Ayrshire Football League
Like (probably, I assume) many of the people participating in this project, I've been on occasion taking a stab at Category:Articles lacking sources from June 2009. Starting from the earliest letters in the ABCs it what gets the most attention, and the first 8-ish or so are waiting at PROD or AFD, but we've hit a block. Ayrshire Football League failed AFD because sources exist at the British Newspaper Archive, but because it's paid, nobody who wants to go about adding the sources has access. Technically speaking its not any more important than any of the other articles, but aesthetically, this is annoying. Anyway, if anyone wants an easy article to ref, and has access to the Newspaper Archive, that would probably be helpful. Mrfoogles (talk) 02:44, 10 March 2025 (UTC)
In the course of looking up a specialized database for the AfD that few others have access to, adding in one of those references would have been more beneficial than just mentioning their existence. Otherwise the article ends up in the same ignoble state it was in originally with little to show from the discussion. SilverserenC06:24, 10 March 2025 (UTC)
Just a tip that I just learned: British Newspaper Archives does have some items that are free to view - and you can limit your search to such items by ticking the "Free To View" box under "Access Type". A search for "Ayrshire Football League" turned up five; I'll add one or two of them to Ayrshire Football League later today when I'm on a decent computer. Cheers, SunloungerFrog (talk) 07:04, 10 March 2025 (UTC)
Try to hit 90 by March 31
Hello 2025 Unreferenced Article Marathon participants. Thanks for your dedication. As of right now, we have 13 editors who have at least 90 points on the 2025 leaderboard (January through March, total).
All editors with 90 points recorded by the end of the day on March 31 will receive barnstars in April, as that's the equivalent of 1 per day over a 3-month period. So if you're close and you have the time...try to get to at least 90 if you can. Cielquiparle (talk) 21:23, 22 March 2025 (UTC)
And don't forget to remove the {{unreferenced}} tag if you've added a reference. (That's how the bot is tracking and assigning points.) You can replace it with another tag like {{refimprove}} or {{more citations needed}} or {{single source}}, or add {{citation needed}} within the article body if you're doing a quick "ref-and-run". (This means you, @LucasJanssen4444!) Cielquiparle (talk) 22:32, 22 March 2025 (UTC)
I didn't realize I had referenced this many articles and now am oddly noting my total is listed as 99. I'm now considering whether I have perhaps been spending more time doing this than I thought. That is a lot of Wikipedia articles. Mrfoogles (talk) 01:09, 25 March 2025 (UTC)
Wow, I had no idea this was going on! So, per above I figured I must've "fixed" 26 articles, but the bot is counting 17. I was just using the sheer change in numbers on that petscan, does this mean 11 of the articles that were originally on that list were probably handled by someone else? Valereee (talk) 13:34, 25 March 2025 (UTC)
That could have happened -- I often look at the category counts and see it's gone down more than I fixed. If you want to really find out, you can probably go back and look at your contributions (although that'll take a while). Mrfoogles (talk) 05:10, 29 March 2025 (UTC)
June 2009 sources
Came here to this talk page to say two things. First, June 2009 is down to 3 articles that aren't under deletion discussions! Teemala, Kalwas, and Stored energy printer. There seem to be a lot of small Indian villages. Second, does anyone have any advice on how to handle the small Indian villages? Many appear to exist but not show up in the census.gov.in Population finder. E.g. Kudkelli -- oh wait, nope, there's a typo in that article name, the source I wasn't sure about lists it as "Kudkeli", and the population finder apparently only finds the exact correct name. In any case, does anyone have any advice on sources other than the population finder for Indian villages? And does it make sense to nominate them for deletion if no sources can be found and they don't appear in the population finder? While some are misspelled in Wikipedia, all the various websites, such as the current source on Kudkelli mentioned above, derive from censuses anyhow. Mrfoogles (talk) 04:57, 29 March 2025 (UTC)
I think that Teemala/Timla likely dates to a pre-2011 census, as neither of those names work on the 2011 pop. finder. Mrfoogles (talk) 05:07, 29 March 2025 (UTC)
I located Teemla/Teemala/Timla based on the description in the article and added the coord to the article and the Open Street Map way ID to Wikidata. Turtlecrown (talk) 14:50, 29 March 2025 (UTC)
Back to the question about Indian villages, I often add a site:gov.in (or the state's url ending) to Google searches, or look for it on the district's website, or search for it together with nearby places on the News tab, which often helps with finding alternate transliterations. And a lot of other steps too, depending on level of desperation. Turtlecrown (talk) 23:15, 29 March 2025 (UTC)
If you can't find it on the Census finder, or other reliable source, it makes perfect sense to just send it to AfD/PROD. Either references turn up, or it's deleted/redirected. ARandomName123 (talk)Ping me!21:35, 30 March 2025 (UTC)
@Lenticel - there will be an alert on the Watchlist pages as we get closer to the drive but @ARandomName123 do you think we could send talk page messages to people who signed up for previous drives? Kazamzam (talk) 14:36, 5 March 2025 (UTC)
I searched the lists in the Progress section for food and drink articles, and didn't find any in any of these, so I checked here and still came up with zero. Which unless someone has been systematically working on that category seems unlikely. Did I screw up the search, or is it actually true there are no food and drink articles that lack sources? Valereee (talk) 12:15, 16 March 2025 (UTC)
I think it wold be possible. Each one would probably need doing individually because it's working on subcategories, and the choice of category depth and starter category affect what is included and left out in rather arbitrary ways, due to factors like different nesting depths and the fact that they are not strictly hierarchical. At least it's easy to see how it works, unlike articletopic, and to tweak it if needed. There's a few already prepared, labelled 'Petscan query (Category)', on this now-archived discussion from last year. (On a side note, it would be interesting to see what progress has been made on those topics! [Edit: it was not interesting]) Turtlecrown (talk) 15:26, 20 March 2025 (UTC)
Appreciate that you did this anyway. We can all appreciate the uninterestingness, together (I wonder why the total # of article in the categories consistently decreases significantly, though -- are articles being deleted, or just recategorized? Categorized more deeply? That is interesting.) Mrfoogles (talk) 03:21, 25 March 2025 (UTC)
Speaking only for myself, I've gone through probably the first 100 of the 518-article depth 5 food and drink petscan. A very large portion weren't actually food related (for instance, a fishing village and dinnerware items got caught), or were tangential (a Canadian food ministry) or just not really in my interest/skill set (commercial producers/products), but for the ones that were actually about food and drink, I've fixed a few articles by finding sources, but I think I've converted to redirects more than anything else. The petscan now has 492 articles, so I've done something with about 25. When I spend a session on the list, I can get through about 30-35 at a time, so I'm thinking after another ten or twelve sessions, I'll come back here and get advice on whether I should do a depth 4 or a depth 6 next. :) I don't really understand category depth very clearly. It's a nice little to-do list for when I feel like this kind of work. Valereee (talk) 13:26, 25 March 2025 (UTC)
Category depth is how many sub-categories of Category:Food and drink the PetScan looks into. For example, at depth 5: Food and drink (0) -> Drinks (1) -> Non-alcoholic drinks (2) -> Tea (3) -> Tea varieties (4) -> Green tea (5) -> Japanese green tea (6). Then repeat this for all subcategories. Which means you also unfortunately get, for example, Food and drink (0) -> Food politics (1) -> Agrarian politics (2) -> Peasant revolts (3) -> Popular revolt in late-medieval Europe (4) -> Tudor rebellions (5) -> Nine Years' War (Ireland) (6). Maybe starting from a food and drink subcategory more specific to your interests would be more useful, but it's never going to be a perfect solution due to how the categories are structured. Another useful tip is to add negative categories (eg Politics) if the same ones keep popping up and 'polluting' your list. Turtlecrown (talk) 08:16, 27 March 2025 (UTC)
@Mrfoogles The reduction in the number of unreferenced articles per category since the last time they were counted is in line with the average reduction in the unreferenced article total (about 27%) since then. I think this is largely due to the work of this group, considering that there was about a 11% net reduction during our last backlog drive, despite us referencing more than that (due to the constant influx). However, articles in these particular categories have actually become a larger proportion of the articles that are tagged as unreferenced. That might be a caused by a greater influx of new articles/tags in popular categories than the average, or by less work being done in these categories (eg due to waning enthusiasm for referencing endless villages and albums). Turtlecrown (talk) 13:33, 25 March 2025 (UTC)
Hello from the french-speaking wikipedia !
The project on the french-speaking side of the Wikipedia world is going very well, and we just discussed yours.
Currently, we only have around 2,000 categorized articles with the template. In fact, we need to track uncategorized articles (is "untemplated" a verb?) that effectively have no sources. Two months ago, our query identified about 200,000 uncategorized articles with a high probability of lacking sources. Now, that number is down to around 160,000, thanks to a nice little bot (a cute turtle) that checks whether these articles in the query have useful items in Wikidata to automatically link them. We manage to process around 1,000 articles manually per month.
We're still trying to find new ways to automate or semi-automate the process. For example, our tech team is planning to test a "soft" sourcing method, where the bot would add references in the article’s Talk page, so verification and integration remain human-driven.
So, when I see that nearly 67,000 articles are categorized here on EnWiki, I'm truly impressed—kudos to all of you! I have a few questions:
Are all unreferenced articles categorized? (By that, I mean: are there any missing templates?)
I ran a quick PetScan and found 30 articles that appear in both your category and ours. How can we help you the best ? Because we’ll have sourced them within about six months at most. I think there's more and my PetScan query is bad. I'll also ask our tech team to check the query—there are probably more articles to be found.
Following my second question, with Battle of Khankala (1994) as an example, i did some "hard" sourcing showing that the old version contained few wrong information. I won't try to translate it into a non-native langage. I applied a trad template... But maybe there's a better way to help you ? Nanoyo88 (talk) 09:21, 27 March 2025 (UTC)
@Nanoyo88, putting the {{expand language}} template on the enwiki article after you have done the hard work finding references in frwiki is very helpful. If you were feeling especially keen, you could also put {{refideas}} on the Talk page; this is a way to actually list out the references you found. Cheers, SunloungerFrog (talk) 09:49, 27 March 2025 (UTC)
First, terminology: I'm going to assume that when you say 'categorized' you are referring to articles tagged with the {{Unreferenced}} template, and therefore added to an unreferenced maintenance category. And 'uncategorized', therefore, means unreferenced articles that don't have a tag (or as you say, 'untemplated'). This might seem like a picky clarification, but on en.wiki we also have the uncategorized taskforce for generally uncategorized articles, which might be what most people first associate with the word. For simplicity, I'll use tagged or untagged.
Second, my questions: Is the fr.wiki project responsible for referencing the fr:WP:Atelier de relecture, or is there a more specific one? And do you need to track untagged articles that effectively have no sources in order to tag them, or for some other reason?
Third, to your questions:
Question 1: Are all unreferenced articles tagged? No. Tags are added by editors by hand or using semi-automated tools, usually not by us. Several hundred tags are added each month. Predecessors of this project seemed more involved in adding the tags, whereas our current WikiProject is mainly involved in resolving them. 111,000 unreferenced articles were found and tagged by a bot in December 2009 (see notes under Historical data).
Question 2: How can the French language referencing project support the English language one? Adding {{expand French}} as you did at Battle of Khankala (1994) is probably the easiest way. It lets us know there is more going on at the French page and that we might want to look for sources there. There are also many other unreferenced English articles that are perhaps easier referenced by someone who can speak French (I often do this!) such as those found in this (again, imperfect) Petscan.
Thanks for the terminology clarification. Let's use tag and untagged. The Wikiproject "Atelier de relecture" does not aim on referencing. It's the same as Wikipedia:Peer review. The french project i'm talking about is Projet:Articles sans sources (also nammed PatASS for "Patrouille des Articles Sans Sources [Unreferenced article patrol]"). You can see here it's quite young but growing. Btw, i see you're using a bot for your stats, it may interress us (ping @Mr Tortue, our "turtle tech"). For the track question, see below.
Q1 - thanks for the data. That's huge ! We also had a long history of tagged article with no real project handling them since november 2023. We have managed to handle all the banners so that the oldest category is now from August 2024. In addition to that, we have untagged articles that we manage to find with a long query (see here), the widest (with few false result) is around 160K. We use a bot on these query to directly find in the wikidata if items can solve the unrefferenced problem (the bot did 40K+ referrencing in a month). I hope you don't have too much untagged articles. So to answer your question on why we track them : because there was no project like yours in 2009 in the french-speaking WP. The unrefferenced french template was only created in 2015 ! So we track them to tag them because we're actually referencing 10x quicker than the "natural tagging" by the community (around 100/150 per month without our tags).
Q2 - Ok. I'm not sure everyone will put the template as i did, but helping with the subject of the petscan could be a good idea. Nanoyo88 (talk) 10:33, 27 March 2025 (UTC)
You probably know this already but PatASS is a good name for a project, if you might not be able to get away with it in English. Mrfoogles (talk) 05:14, 29 March 2025 (UTC)
We have a template "Liens" that automatically put links from wikidata based on items linked (such as dictionnaries, encyclopedia, thematic ressources, authority, etc.) See for example in french Osei Tutu I. In the bottom, you can see "Liens externes" which links to BlackPast, Britannica, etc. All those links are from the template "Liens". The bot is just looking in the query of untagged articles without the template "Liens" if they have a usefull item in wikidata for referencing. If yes, it simply add the section + template... And voilà :) Nanoyo88 (talk) 09:34, 31 March 2025 (UTC)
I just graphed the number of articles in June 2009 over time -- see File:June 2009 unreferenced articles graph.png, and there's a fairly constant slope until about February 2025, when the slope (rate of referencing) massively increases. This is the exact time when it became the oldest unreferenced articles category -- May 2009 was deleted on January 31st. So, it looks like people really do focus on the oldest category. Mrfoogles (talk) 01:18, 30 March 2025 (UTC)
Actually, that graph has an error. All the data points after the first one should be shifted two months to the right. But it's mostly the same. Mrfoogles (talk) 01:26, 30 March 2025 (UTC)
Interesting. I tend to focus on the oldest categories (month/year) but then sometimes choose by subject too, as I think you need to also keep the search for references interesting to you. Great to see all this progress being made. Coldupnorth (talk) 07:19, 31 March 2025 (UTC)
Top 10 URA leaderboard for Mar 2025
Well done to all who contributed citations to articles tagged as Unreferenced last month. Here are the editors who made it into the Top 10 in March 2025:
@Silver seren continued to dominate the leaderboard in March, but @Cakelot1 was not far behind. @Turtlecrown came in a respectable third, with a strong performance for the second month in a row. In an exciting development, @Mrfoogles debuted with a leapfrog into fourth place, while @Coldupnorth bested February by joining the ranks of the monthly 100+ pointers in March.
To track where we are for the year to date, see the WikiProject Unreferenced articles leaderboard for 2025. You have to add references to at least 5 different articles before your stats will show up in the yearly view; however, if you click on "Apr" you'll see the full list of editors who have been active this month, with links to their edits.
And I do still mean to distribute barn stars this weekend to the 14 editors who hit 90 by March 31 (equivalent of 1/day in the first three months of the year). Cielquiparle (talk) 09:44, 3 April 2025 (UTC)
Hope everyone is doing well. Many of us have probably seen the banners for WikiConference North America 2025, which is going to be held in New York in October. Since this project seems to be bucking the trend of declining WikiProject activity and has made substantial progress on our target goals (overall backlog clearing, unreferenced BLPs in particular) and some high-visibility successes (turning unreferenced stubs into DYKs), would anyone be interested in pitching some kind of discussion/talk to the WCNA organizers? I've never been to a conference but I think I'll be able to make this one and if we continue at the good clip we've been maintaining, I think it would be cool to spotlight our successes, discuss challenges (paywalls, language barriers), and maybe get some new members. If anyone has either the interest or the experience to get involved, I would much rather do this with someone else than fly solo. And if anyone has been to one of these events before, please share your vast wisdom. Cheers, Kazamzam (talk) 15:02, 11 March 2025 (UTC)
Wow. What a month. Well done to all who added citations to articles in April. Your contributions helped us get the total number of Wikipedia articles tagged as Unreferenced under 64k.
Why are there 11 editors listed in the Top 10, you ask? First of all, it was nearly a tie for 10th place. Secondly, for anyone following the competition in April, it was hard to ignore that things got a little...spicy.
There were a whopping 7 editors with 100+ points. Those were fighting words indeed from Someonefighter, who topped the leaderboard several times but just could not hold onto the lead for very long, as the mighty Silver seren kept clawing back and finally blew past everyone at 500+, a monthly URA record for 2025. Between these two editors with the initial "S", they logged 952 points in April. Simply staggering. (The next best first-initial cohort, the three editors whose names start with "C", including Cakelot1 in fourth place, Coldupnorth in seventh place, and yours truly, delivered an aggregated 419 points.)
New joiner LastJabberwocky worked all the way up to an impressive third place finish, while Thefallguy2025 who joined the action in March jumped to eighth place in April. Also pleased to see many familiar editor names up and down the board, including JoeNMLC who placed sixth for the month, but remains in fourth place overall for the year.
What's next? There is a lot going on in May. Lots of ways to get involved, and lots of opportunity to climb the leaderboard while other editors are doing other things. :) Remember, you need at least 5 points for May for your results to be displayed in the main yearly leaderboard (but you can track everyone's progress in detail if you click on "May").
A few insider tips for URA fans:
Log at least 61 points total for April and May 2025 and you'll receive a barnstar (as that's the equivalent of 1 citation a day in Q2 minus the backlog drive).
It's been a great deal of fun working on this project, I went far and hit a little bit of a roadblock: I've been through the entire books category, and have added citation to every book that exists in https://isbndb.com/ , the self proclaimed "biggest book database in the world"
@Cielquiparle I don't know if you have trackers for specific categories, but I've managed to get the books category from around 680+ articles to about 230+ articles. The remaining are going to be more time consuming, as some only exist in WorldCat, and some don't even exist there. I don't know if the books that aren't listed there are ever going to get citations. Additionally, I've stumbled upon many articles that are misplaced (for example, articles about authors, TV shows and other unrelated stuff). If anyone knows how to fix this, please tell me and I'll go over the category again and fix categorization. My goal in the project was to get the books category to 0, since I love reading books.
Honestly, I joined this project when I was a very new editor, and it taught me a lot about the essence of wikipedia. Specifically: verifiability and what sources are considered good. This helped me by a great deal now that I'm transitioning into editing articles in topics I am more knowledgeable about. But fear not! As long as I edit wikipedia, I will be a major participant in this project
p.s: I planned to make about 150 citations from worldcat and win the competition. But due to recent events, I'm locked to a VPN network, which worldcat blocks. I hope someone can continue my mission and finish the books category! Please contact me if you know about other big reliable databases Someonefighter (talk) 08:50, 2 May 2025 (UTC)
@Cielquiparle - one thing that does irk me about the leaderboard is that it is, as far as I'm aware, based on the removal of existing tags from article; this excludes when editors add articles to untagged, unreferenced articles which I have found quite a few of (biographical articles of medieval Japanese women are plagued by this, for some reason). Not quibbling over points - frankly I am thrilled to see this project stand on its own without me being the only person posting on the talk page - but just noting that it's an area that is overlooked. We've discussed the issue of getting an "accurate" count of unreferenced articles but it is quite a challenge.
Cheers to the competition this month and the upcoming drive! The next update will be on 4 May. See you soon! Kazamzam (talk) 13:44, 2 May 2025 (UTC)
At least as someone that's a part of this effort, you could purposefully add the unreferenced tag in one edit and then remove it in the next (while adding the reference you were planning to add in the removal edit). That should make it be counted properly. Doesn't help for those not a part of this group though, unfortunately. SilverserenC16:05, 2 May 2025 (UTC)
I honestly didn’t expect to end up in the Top 10, but I’m really grateful to be part of this project. Working on unreferenced articles has taught me a lot about sourcing and the importance of verifiability, and it’s been surprisingly rewarding to see the difference even small edits can make. Huge respect to everyone who contributed this month — especially Silver seren and Someonefighter, your work was incredible. I’m looking forward to continuing in May and doing my bit to help chip away at the backlog. — Thefallguy2025
Thanks for the update! Absolutely incredible work by Silverseren and Somonefighter. That's a combined nearly 1000 tags between just you two! ARandomName123 (talk)Ping me!16:54, 2 May 2025 (UTC)
May 2025 update
Happy spring, fellow editors in arms! I hope everyone is doing well and May the Fourth be with you.
Headline: We cleared 5,733 articles and are now hovering just over 63,000 (63,303 to be exact)! My predictions for clearing 50,000 by the end of the year seem increasingly likely. For yourself and your fellow editors, please clap.
Minutiae: For anyone interested in a more detailed breakdown of the numbers - average was 69.1 articles; median 23; mode 20. This is a significant jump from last month (average 35.3; median 12; mode 11), so really well done! Of the categories being tracked, all but ONE had a decline of at least 10 articles and over 3% of their starting number compared to previous update (I consider this a growth metric that we're being comprehensive in clearing categories across the board), so this is fantastic. The smallest decrease was December 2019, which only decreased by 9 articles.
Highlights: May and June 2009 are in dustbin of history! The leaderboard also seems to be bringing out the competitive spirit in people, which is perfect as we get ready for our June 2025 drive.
Low-hanging fruit: The infamous September 2019 is a runty 81 articles, dangling precariously like a sinner in the hands of an angry God. If you're not a fan of square numbers, June 2024 is a bite-sized 94. Food for thought.
High-hanging fruit: Everyone's favourite BFC (Big Friendly Category), December 2009, is a mouth-watering 8,278 articles as of this writing, after a decrease of 830. The other high-hanging fruit are, still, the Frustrating Five (name open for revision): January 2013 (924), May 2019 (1,639), June 2019 (3,604), September 2020 (961), and March 2024 (854). Eagle-eyed readers may have caught that previous squad member April 2019 has been ousted from the hall of infamy, so well done there. Once again, September 2020 had the lowest percentage of change between updates (1.43%). Godspeed to anyone working on these.
New challenge: We're back with the ties! December 2015 and January 2016 are in a dead heat of 248. If you strongly prefer one year over another, pick a side and hold the line.
Announcements: June 2025 is right around the corner! We seem to have found the rhythm of how the drives work, which is great, and additional organizational help would be much appreciated by new members. As always, I think the goals are a) clearing 10,000 articles and b) clearing the entire backlog (whatever we have) of unreferenced BLPs, but happy to have secret, member-specific goals. All the best, Kazamzam (talk) 00:08, 5 May 2025 (UTC)
Everyone making great progress, well done all! It looks like User:Someonefighter may have been banned unfortunately so we may all have to work that bit harder to reference those articles this month! Coldupnorth (talk) 08:49, 8 May 2025 (UTC)
Why are almost all Indian villages spelled incorrectly?
Someonefighter and I are struggling through this area of unreferenced articles right now and it is a pain. The issue is that nearly every Indian village article on Wikipedia is named incorrectly. Often by just a letter or two or phonetically (rather than -u, it's -oo), but I still just don't understand why so many of them are wrong. Do these places not have actual proper spellings? The census I'm working from definitely has a spelling they could have used.
This feels like when I try and work on a biographical article from 1800s America and there's no consistent spelling of the person's last name, likely because the person was illiterate and just told the census takers and journalists over the years their name and each of those people just assumed the spelling in different ways based on what they heard.
Is that what's going on here? Indian villages just didn't have consistent spelling until recently? Or is it just that the many different editors that made these articles are terrible at spelling (or at least terrible at looking up how to spell these villages properly)? SilverserenC01:39, 11 May 2025 (UTC)
@Silver seren:Hunterian transliteration is my best guess. From that article: 'Opponents of the grapheme transliteration model continued to mount unsuccessful attempts at reversing government policy until the turn of the century, with one critic calling appealing to "the Indian Government to give up the whole attempt at scientific (i.e. Hunterian) transliteration, and decide once and for all in favour of a return to the old phonetic spelling."'. I don't envy that part of the unreferenced queue. --Engineerchange (talk) 03:08, 11 May 2025 (UTC)
Well, the two of us are definitely making progress. Though I'm skipping any of the villages that I can't find anything in the census for even under reasonable alternative spellings. So we're not getting rid of all of them. It really feels like that the longer the village name (and some of these are super long), the more incorrect the spelling is. SilverserenC03:13, 11 May 2025 (UTC)
Silver seren I think it is likely that Indian villages are often not spelled consistently in English. It is also true that people who wrote completely unreferenced articles are likely to have been terrible at looking stuff up by current Wikipedia standards. I found an interesting book on Google books Mapping Place Names of India (Anu Kapur, 2019) which says The purpose [of transliteration] was clearly to serve the English speaker and reader. Independent India has not been able to put forward an 'Indianized' system of transliteration for 70 years! and goes on to discuss the issues of Standardization the Anglicized way.
It is useful to try and find sources and standardise spellings, but not surprising that the original editors did not do so.
As an example, Someonefighter has recently moved Osiyan, Unnao to Aosiyan. I can find Google book hits for Osiyan, Unnao, but not Aosiyan, Unnao. The article mentions J.D.V.M. Inter College, where a Google search for the college finds Instagram and Facebook accounts saying osiyan and Yayschool saying Aosiyan. Villages may have different local spellings in Hindi and other local languages: according to file:Languages of Uttar Pradesh State.pngUnnao is in the area where it is difficult to say whether it is Awadhi or Kannauji. TSventon (talk) 09:02, 12 May 2025 (UTC)
Update: it took me a while but I eliminated almost all villages/towns in India, as well as almost all in Pakistan and some of Sri Lanka (all villages of Sri Lanka). Back in April (when I first looked it at) south Asia category was about 2.5k articles. Now it is 615. I am currently looking for a list of towns in Sri Lanka that isn't behind a paywall (I have found the villages one used it, but it does not contain towns)
I assume you're running a script for edits, Someonefighter? Since it looks like you did several hundred in just a few minutes and did the template removals in a batch afterwards. Be careful with that, since you don't want to run afoul of WP:MEATBOT related policies. SilverserenC23:09, 13 May 2025 (UTC)
Thank you for your concerns, @Silver seren. I prepare all pages in a lengthy process (in the case of citation), then submit them all at the same time (alt+tab -> alt+shift+s many times). For this reason my edits are mostly fractured, since it is more convenient to split them into multiple actions. I do not use scripts to directly edit or submit the articles, only to determine if they're in the lists as I've said earlier. Someonefighter (talk) 08:54, 14 May 2025 (UTC)
June 2025 Backlog Drive
OK informal backlog drive planning committee, I know everyone is busy but we have a backlog drive to run in June. What do we need to do to get the page out of "Planning" mode, @ARandomName123? A little concerned that only one other person has signed up for the backlog drive so far. (Could we maybe mimic what the NPP Backlog Drive folks do...banner at the top of the backlog drive page asking for signups?) I also kind of like their registration box but whatever, we just need people to sign up at this point. @DreamRimmer At what point is it optimal to run a Watchlist page notice thing? Pinging @Kazamzam @SunloungerFrog. Any other volunteers who more or less know the drill...? (I actually don't want to stop the goodness that is the marathon so please don't stop running, as we still need to add references in May as well.) Cielquiparle (talk) 02:20, 12 May 2025 (UTC)
We just have a particular focus/push to reference as many articles as we can. You probably don't need to change anything you're doing, @Someonefighter, given your position on the leaderboard :) apart from include #JUN25 in your edit summaries so that they get counted. Cheers, SunloungerFrog (talk) 07:51, 12 May 2025 (UTC)
I wouldn't worry too much about the current amount of sign ups yet. I don't believe we've done any advertising whatsoever, and I'm pretty sure most comes from the watchlist notice anyways.
As for getting out of planning mode, I there isn't much left to do on the drive page itself. Maybe spread the word to the usual places, put out a talk page notice, WP:CBB, etc. ARandomName123 (talk)Ping me!14:23, 12 May 2025 (UTC)
@Cielquiparle - so sorry for my prolonged absence, this month has been crazy busy. I think the drive planning to increase sign-ups so far (banner, watchlist header, talk page + project page, talk page reminders for individual signups) has been great. I agree with @ARandomName123 that not much more is necessary but it would be cool to get the word out to other WikiProjects that people are active in - I can do WP Japan for example. I think we could be more detailed in the talk page template about why this is important (age of the backlog, for example).
Also, and this is just me, but I think we should have a goal once again of 10,000 articles. Consistency with previous drives is a plus, but I'm open to other arguments. Alternatively, we could switch it to 5,000 articles because we are almost certain to blow through that and it would be a boost to editors to set and meet goals. Cheers, Kazamzam (talk) 14:00, 21 May 2025 (UTC)
@Kazamzam I went back and checked, and we had a lower target for November 2024 in part based on the number of signups. Personally I like that you set us the 50k goal for end of year, which we seem well on track to achieve, plus it's easy to remember – an epic milestone for URA if we do achieve it by mid-year. Cielquiparle (talk) 16:55, 21 May 2025 (UTC)
@Cielquiparle - complaining about shifting goalposts when I'm the one shifting them...I will go to atone for my sins under a roaring waterfall like a Shugendo yamabushi (nice catch, thanks!) Kazamzam (talk) 17:27, 21 May 2025 (UTC)
@Kazamzam@Cielquiparle Have we reached out to project who have a lot of articles in our list? Off the top of my head I have Hinduism (many temples) and songs/music as well as sports. There are certainly many more that we could potentially attract editors from Someonefighter (talk) 18:33, 21 May 2025 (UTC)
@Someonefighter Great idea! Please go ahead and reach out. I think we should each reach out where we feel we can advocate why URA and the backlog drive should be of particular interest to whichever WikiProject. Cielquiparle (talk) 21:33, 21 May 2025 (UTC)
@Cielquiparle I have made this notice (basing it on your one). I will add it tomorrow to relevant pages. feel free to modify it and use it until then.
Please consider signing up for the June 2025 Unreferenced Articles backlog drive.
This project has over (insert number of articles here) recorded unreferenced article. Help us improve your project by adding citations and participating in the June push.
@Cielquiparle, I'm dusting off my Petscan and Massviews searches for the popular unreferenced articles section. I don't really know about any other way of drumming up support, other than sending a talk page message to those who have subscribed in the past. Is it inappropriate to post on other groups' talk pages if we are part of those groups? For instance, I would happily post a short message on new pages patrol talk provided I won't get tutted at. Cheers, SunloungerFrog (talk) 05:30, 20 May 2025 (UTC)
@SunloungerFrog That's a great idea. Quite certain others have done so in the past. Might even be helpful to explain to them how to use Petscan to find the Unreferenced articles relevant to their respective topics. Anyone else up for reaching out to other WikiProjects? Cielquiparle (talk) 11:07, 20 May 2025 (UTC)
I've updated the message from last time, over at Wikipedia:WikiProject Unreferenced articles/Newsletter. Please feel free to make any adjustments you see fit, and ping DreamRimmer once it's ready (I'll do so tmrw). You can see it below
Extended content
== Unreferenced articles November 2025 backlog drive ==
There is a substantial backlog of unsourced articles on Wikipedia, and we need your help! The purpose of this drive is to add sources to these unsourced articles and make a meaningful impact.
Barnstars will be awarded based on the number of articles cited.
Remember to tag your edit summary with #NOV25, both to advertise the event and tally the points later using Hashtag Summary Search.
@ARandomName123: Looks great! We are up to 33 participants now. And even if a few of the people who receive the mass message have already signed up, they have Talk page followers who will see the message and sign up that way, probably. Cielquiparle (talk) 03:41, 21 May 2025 (UTC)
I was just wondering, since I saw some mention above about using this tag for the edits to "count". Is that really necessary since we have the leaderboard now? (And we'll have a June specific page for it when that month starts.) So, any removal of the template during that month will show up on that leaderboard, with diffs even for the removal. Why would tagging your edit with #JUN25 be necessary? You can still include an edit summary mentioning this project, if promotion is the purpose. SilverserenC23:45, 21 May 2025 (UTC)
@Silver seren @ARandomName123 It also prevents editors from claiming points if all they did was remove the tag (e.g. when another editor added a citation but forgot to remove the tag). All #JUN25 tagged edits are subject to review by other editors (who get 0.5 points for reviewing) and points can be deducted if it turns out that the citation added was inappropriate, or if a citation wasn't added in the first place (even if it's for a very good reason). (We do however allow dummy edits after the fact to add #JUN25 if someone forgot to add it in the same step as adding the citation.) The two leaderboards will start to diverge for these reasons in terms of points, but my hypothesis is that they will end up being quite similar in terms of ranking. Cielquiparle (talk) 03:27, 22 May 2025 (UTC)
On second thought, the two leaderboards may differ quite a bit because editors can accumulate a large number of points for reviewing, which actually is a key part of the backlog drive. Cielquiparle (talk) 04:12, 22 May 2025 (UTC)
@Kazamzam The main page. I kind of like the idea of clicking a single button that instantly takes me to a random completely unreferenced page where I can learn about something new and improve the article as well, as opposed to going to a list and choosing a certain topic/page that you are already familiar with. GOAT Bones231012 (talk) 14:44, 21 May 2025 (UTC)
@DreamRimmer - what do you think of this? I don't think it should be at the top of the main page but maybe somewhere in the header would be doable and convenient. @GOAT Bones231012, fair warning you will find this to be a LOT of train stations and villages in the Indian subcontinent. Kazamzam (talk) 14:48, 21 May 2025 (UTC)
I figured lol. Unseo station was one of the first ones that popped up for me yesterday. The user can always refresh the button and be taken to a different random page, but if that’s what most of the unreferenced pages are, maybe it’s not such a bad thing to take some of the backlog off these topics? GOAT Bones231012 (talk) 15:06, 21 May 2025 (UTC)
@GOAT Bones231012: No comment on adding this to the main page. If you are interested, you can add the following code to your common.js. It will add a link to your tools menu, and clicking it will take you to a random unsourced page each time.
mw.loader.using(['mediawiki.util'],function(){mw.util.addPortletLink('p-cactions','/wiki/Special:RandomInCategory/All_articles_lacking_sources','Fix a random page lacking sources','t-random-unsourced','Go to a random unsourced article');});
Well, you can debate whether it should be there, but the purpose (which I generally agree with) is to kind of give credit to the WP:URA project for putting in the article's first reference. The goal is not to mark articles of interest to the project, but to take credit for improving articles so that people know about the project that is citing the articles they're reading. This helps the project survive (by ideally providing some influx of new members/people who are at least aware of it) and also makes people who provide an article's first reference feel good because they get the reward of leaving a nice template after they improve an article. The idea is that, even if the article does get majorly expanded afterwards, the addition of the first reference is still significant enough to leave a template. For most articles referenced by this project, they don't get more references independently afterward, so a significant portion of their quality is due to this project.
So no, if it ceased to be displayed it would go against the purpose of the banner, which is to make the work the project does visible. It's a different kind of thing. Mrfoogles (talk) 22:49, 31 May 2025 (UTC)
Remember to add #JUN25 to edit summaries
To all you hard-working 2025 Unreferenced article marathon runners out there: PLEASE PLEASE PLEASE just for the month of June, remember to add #JUN25 exactly once to eacn article you add citations to, which was previously unreferenced. This is what is required to appear on the special June 2025 backlog drive leaderboard which appears on the main WP:JUN25 backlog drive page. This means you, @NorthernWinds. Cielquiparle (talk) 07:56, 1 June 2025 (UTC)
@Clarityfiend Thanks for merging the content, but no it doesn't count. At first, I was going to suggest that if you really want a point for the purposes of the backlog drive, you could add a source to the unreferenced article (with the edit summary), THEN merge the content to the target page...but the rules on the main backlog drive page clearly say that "blank and redirect" doesn't count. Neither do PRODs or AfDs. Cielquiparle (talk) 04:00, 2 June 2025 (UTC)
Backlog drive schedule
Hi all—letting you know that I have created WP:DRIVES, to help drive schedulers avoid overlapping drives. When you schedule the next drive, feel free to make use of it, and it would be greatly appreciated if you could add your future drives :) HouseBlaster (talk • he/they)23:45, 8 June 2025 (UTC)
The problem with backlog drives is that although the surge in participation is exhilirating while it lasts...activity tends to fall off immediately afterwards. In the case of WikiProject Unreferenced articles, however, we decided we wanted to create a "Marathon" edition of the competition that was running year-round. Thanks to @ARandomName123, there is a yearly leaderboard you can track here.
Why does it look different from the leaderboard you have been looking at daily at WP:JUN25? It's because it simply counts edits in which "Unreferenced" tags are removed. (@Clarityfiend: This may be of interest to you.) Also, users only appear on the yearly leaderboard once they have reached 5 points in a single month.
Collectively, May 2025 turned out to be our strongest month yet, thanks to the dedication, competitiveness, and research and citation skills of many editors. Turning now to the Top 10 leaderboard:
I have to confess that I was startled by the final outcome, as I could have sworn NorthernWinds (formerly know as Someonefighter) had managed to top Silver seren at the end of May...but sure enough, the mighty Silver seren really doesn't like to lose. The same top two editors from April 2025 took activity to a completely new level in May, both reaching an unprecedented four digits in a single month. At the same time, it was uplifting to realize that they were collaborating as well, as they powered through adding citations to a large number of unreferenced articles about Indian villages. (See section above.) Huge applause for this prolific pair.
Cakelot1 emerged as a force to be reckoned with, in third place with 454 points. There was a considerable delta between third and fourth place, with JoeNMLC, Coldupnorth, and yours truly in the 100–200 range. As for the rest of the board, it was good to see LastJabberwocky return with another strong showing. WikiProject URA and June backlog drive organizers ARandomName123 and Kazamzam emerged from their exams to place in 8th and 9th place, respectively, while Turtlecrown managed to take 10th place for the second month in a row.
Many thanks to all editors who added citations to articles during the first five months of 2025. You helped to reduce the backlog by 12,000+ articles since the beginning of the year, even before the June 2025 backlog drive got underway. Yes, next month's edition will look quite different...but with more editors on board to help, perhaps it's a good change of pace as we take some time to review each other's citations and check out the "Popular articles" refreshed several times daily by SunloungerFrog...and continue to explore ways of filtering and finding unreferenced articles of interest using tools like Petscan, Massviews, and Bambots.
Thanks! I was indeed 50 points from #1 of the year at peak. Though recently university has been rough so I am slowing down a little bit (to say the least). Inshallah I'll be back after I'm done with all the tests and projects to trim a few more thousands from the backlog! NorthernWinds (talk) 20:17, 6 June 2025 (UTC)
Well, I can definitely say neither of us will be accomplishing that again. It only worked out because of the massive amount of village articles that could use the same source. Now we've only got the slog left of all the unconnected articles. I am really looking forward to Kazamzam's monthly overview, since I think a good amount of those village articles were in the Frustrating Five categories, so we should hopefully see some major movement on things from the past month. SilverserenC21:42, 6 June 2025 (UTC)
@Silver seren - the Frustrating Five will be waiting for you! If you'd like a blast from the past, here is what the update was from almost exactly 2 years ago when the Frustrating Three (can you believe it) were just making waves, plus June 10th of 2022 was when I first proposed a competition to an audience of crickets. It only took the better part of two years to make it happen and now you and @NorthernWinds clear more articles together than we used to do in two months! I've definitely become less active because of personal obligations, which sucks, but also because it feels like this project has enough momentum to keep moving without me that I can take a bit of a break and let other people be kickass referencers. I'm so proud of everyone on this page and really excited to see how this drive shakes out. Truly thank you all so much! Cheers, Kazamzam (talk) 02:28, 7 June 2025 (UTC)
It looks like the drive is going to quite handily get below 50k at this rate. We only needed a daily rate of about 280 articles and we've been easily surpassing that thus far. SilverserenC02:34, 7 June 2025 (UTC)
Don't jinx it. All it takes is for someone to dump a truckload of new Unreferenced article tags, rightly or wrongly, into June 2025 for the number to start climbing back up again despite our best efforts. (It happened during the November 2024 drive.) Cielquiparle (talk) 04:46, 7 June 2025 (UTC)
One of my frustrations is the amount of energy we collectively spend on AfD discussions where tons of references are identified and then never added to the articles which end up being kept.
For the purposes of URA, is there any way to easily identify Unreferenced articles that have been nominated for AfD in the past?
(OK, during this backlog drive I have found myself doing gymnastics to hunt for sources that were sitting right there in an old AfD discussion all along... I guess it goes to show that looking at the article history and Talk page can save a lot of time, but I was wondering if there is a way to identify them as a group.) Pinging: @SunloungerFrog @ARandomName123Cielquiparle (talk) 05:09, 15 June 2025 (UTC)
@Cielquiparle, try this petscan. It is basically All examples lacking sources in Categories (first tab), then Old AfD multi and Old XfD multi in Has any of these templates: (third tab) with the talk page ticked. If the AfD was closed correctly, then the talk page should always have one of those templates on it. Quick sample of the results seems to indicate that it does the trick. That is going to be my "popular articles of the day" search now! Cheers, SunloungerFrog (talk) 06:21, 15 June 2025 (UTC)
That's awesome @SunloungerFrog. Some of the old AfD discussions have made me laugh. They don't all have links to reliable sources and some of the articles are a bit cringe but I seriously think a lot of improvement will come from this. (It's always amazing with these unreferenced articles when you find the gem beneath the cruft.) Cielquiparle (talk) 08:01, 15 June 2025 (UTC)
Really interesting query, thanks for sharing! And wow, I know standards have changed, but mostly what I'm seeing so far are some of the worst AFD keeps I've ever seen; just walls of no-text votes or arguments straight out of WP:ATA. The past is a foreign country, etc. Suriname0 (talk) 14:11, 17 June 2025 (UTC)
I do my best after an AfD I've been involved in closes to remember to put the sources found on the talk page, so they're at least readily available for future usage. SilverserenC21:29, 17 June 2025 (UTC)
Can you get points for an article if you cite it and then merge it?
@Mrfoogles: Seeing as the article no longer exists, I think adding a reference in this case is pretty much equivalent to adding a reference to an already cited article, so I don't believe you would get points (though the hashtag tool probably counted it). This was mentioned a bit up at #Do mergers count for the drive?. I initially thought it would, but Cielquiparle's reasoning makes sense to me. ARandomName123 (talk)Ping me!16:11, 19 June 2025 (UTC)
Thinking this through another way: If you hadn't added a citation to the article, there would have technically been "nothing to merge". So I'm potentially changing my mind! "Blank and redirect" doesn't count (as the rules currently say), but "add a citation, merge, and redirect" should be OK! Similarly, it's OK to add a citation and then nominate it yourself for deletion or PROD. You did technically add a citation and it matters because it informs the debate and decision making afterwards. Cielquiparle (talk) 16:18, 19 June 2025 (UTC)
Search engine I found
Wanted to drop Marginalia here because based on a few tests (found some sources I didn't know existed on obscure topics) it's really good at finding stuff on those obscure things that only get SEO spam results. I am interested to see what it can do on obscure villages but haven't tested them yet. Not very widely known but seems perfect for this project. Mrfoogles (talk) 18:26, 19 June 2025 (UTC)
Template:WikiProject Unreferenced articles
I noticed that an editor just removed this template from the talk page of an article I'd referenced in the NOV24 drive on the basis that "the article has sources". I know that this populates Category:Articles improved by WikiProject Unreferenced articles, although this only apparently has some 2,800 or so articles in it, and obviously we have done far more than that in this drive alone. I wondered whether we should consider retiring the template and maybe the corresponding category, as it seems clear that they are not used much, and sometimes misunderstood. Cheers, SunloungerFrog (talk) 15:45, 20 June 2025 (UTC)
Hello. It seems you have been considering bots. As a member of Wikipedia:WikiProject Climate change I would like to automatically tag any uncited articles related to our project which are not already tagged. Might you have time to make such a bot? For example as well as looking at the projects on the talk page it might also allow a project to specify a number of strings - say “climate change”, “fossil fuel”, “global warming” in our case.
Alternatively could you just run a bot to tag everything and leave us to figure out what is interesting to us?
@Chidgk1: Hi, a bot to automatically tag pages would require an WP:BRFA, and probably wider consensus then currently here. I can just create a list of pages that are uncited and untagged, then they could just be manually tagged? That's what we did for uncited BLPs, see the list at User:ARandomName123/BLP. If there are a lot, we would prefer if you trickled in the tags, but given there's about 4,623 total articles in your wikiproject, it should be fine to just dump them all (preferably after the drive though). ARandomName123 (talk)Ping me!19:53, 25 June 2025 (UTC)
Using the criteria listed at my comment here, the more stringent scan gives only three pages (Bali Communiqué, Climate change acronyms, Climate change in the Americas), the last two of which I doubt need citations anyways. These three represent roughly 0.07% of the articles in Wikiproject CC, which is similar to the rate I got for BLPs (0.08%).
I also ran the less stringent search mentioned in that comment (only counts inline refs), and there are 61 for your wikiproject. ARandomName123 (talk)Ping me!22:21, 25 June 2025 (UTC)
Sorry for the late reply. I've been away from my computer the last couple days, but should be returning tomorrow. I'll post the list then. ARandomName123 (talk)Ping me!14:36, 30 June 2025 (UTC)
Top 21 URA leaderboard for June 2025 (Marathon rules)
Congratulations and thanks to everyone who contributed to the WP:JUN25 Unreferenced Article Backlog Drive. The goal of reducing the backlog of tagged articles to 50,000 was achieved on June 30, through a heroic final-day penultimate-day effort which saw 715 articles gain citations.
The Reviews section of the backlog drive is technically still in progress – participants have until end of day on July 7 to accumulate more points by completing Reviews (and avoid point deductions by addressing issues identified by Reviewers).
However, for those of you who are now addicted to adding citations to Unreferenced articles – the action does not stop. You can keep going as part of the URA 2025 Marathon, which requires no special edit summaries or signups...the bot is simply counting every instance of an editor removing/replacing an {{Unreferenced}} tag.
You can track the yearly leaderboard here. Please note, editors' names do not show up on the yearly leaderboard until you reach 5 points in a single month. (But if you click on "July", you'll see the detailed view with every single editors' diffs.)
Great to see the strong momentum kicking off the month of July...and in the meantime please also see the Top 21 URA leaderboard for June 2025 under Marathon rules...which includes all editors with at least 100 points.
Great work on the India-specific backlog. I think you'll find in the actual June 2025 Unreferenced Articles backlog drive that @JTtheOG is still in the lead...and anything could change until the Reviews stage closes. ;) Impressive work indeed by @Frost and @Davidindia. Cielquiparle (talk) 19:43, 1 July 2025 (UTC)
Thanks to everyone who got their name in the leaderboard and others who didn’t get their name here. Thanks again for improving our Wikipedia project. TheSlumPanda (talk) 19:55, 1 July 2025 (UTC)
Looks like you added citations to more than 200 articles in June (per WP:JUN25)...but you must have forgotten to remove the {{Unreferenced}} (which is what the Marathon counter counts). Hence the discrepancy between the two numbers. Cielquiparle (talk) 20:22, 1 July 2025 (UTC)
Ah. And so that is a good tip for the Marathon – always add the {{Unreferenced}} tag yourself before untagging it. I do think there is some value to that (beyond personal points) because it informs future analysis in the aggregate. Cielquiparle (talk) 01:52, 2 July 2025 (UTC)
Ah just released that it's just counting removal of the unreferenced tag. I found loads of articles without references or tag. Secretlondon (talk) 20:24, 1 July 2025 (UTC)
By the way, the drive main page says that it will not end "until July 1, 23:59 UTC", that is, in about three hours. AwerDiWeGo (talk) 20:36, 1 July 2025 (UTC)
Mistake!!! In the past I have always been milking it to the bitter end. Not sure how I messed that up but thanks to everyone kept going on July 1 and hope you stick around for the rest of July as well (and the rest of 2025 and/or until we eliminate the backlog). Cielquiparle (talk) 01:57, 2 July 2025 (UTC)
It's kind of incredible that 21 people managed to get over 100 points, even if that isn't exclusively just referencing articles. In every month past this year that number was 2 to 7 people, depending on the month, getting to 100. So that's a massive difference. SilverserenC21:38, 1 July 2025 (UTC)
This has been a great experience. My exp in wikipedia has been limited to just writing articles I know about, template error clean-ups, and adding references where I can. I honestly thought that if I could do 100 edits during this run, it will be good enough, but I am motivated by everyone who is ahead of me, making me want to do better. I am glad I took the leap of faith and join this drive. Thank you for organising, and to participants, please give yourself huge pat in the back! Lulusword(talk)23:51, 1 July 2025 (UTC)
Seconding this! I was never short of conversation topics for the last month because I was learning something new almost every day about some weird and obscure but usually interesting topic. I gained a lot of experience in source-finding outside of my usual areas of expertise, too. Thanks for organizing it, and it was exciting to be part of a project with so many eager participants! Local Internet User (talk) 21:54, 4 July 2025 (UTC)
Before we close out the backlog drive completely... @ARandomName123 I don't think the Reviews have been tallied correctly by the bot. For example, I reviewed a lot more than 5 articles. Cielquiparle (talk) 18:12, 8 July 2025 (UTC)
Hey @Voidxor. The June 2025 backlog drive and the associated review period are officially over.
If you're finding that you miss the referencing action, join the July 2025 marathon. It's rather quiet this month, so those of us who are adding citations are finding it rather easy to climb the leaderboard.
@Cielquiparle: Yes, I saw in the rules that reviews after July 7 don't earn points anymore. So I suppose you're saying I should stop reviewing. How is it that it is "over" though? There are still hundreds of referencing edits that need to be reviewed. I remember in last year's drive that it took two weeks for the tallying to be completed and the awards distributed. See here. — voidxor22:58, 13 July 2025 (UTC)
@Voidxor The tallying has been completed. If you look closely it even says at the top of the June 2025 Backlog Drive and Reviews pages that they have closed and that you're not supposed to alter those pages anymore. (Depending on your settings you might have noticed that those pages have "changed color" as well.) You are welcome to review as many of the referencing edits as you want, but if so please do it as part of the normal course of editing (outside any competition), and please don't keep updating that Reviews page. Awards will get distributed when our volunteers have time. Cielquiparle (talk) 02:56, 14 July 2025 (UTC)
@Cielquiparle: Gotcha! I suppose my settings kept me from seeing the color change, and I didn't see the closing announcement had been inserted at the top of the instructions until after you pinged me. (Doh!) Thanks again! — voidxor15:36, 14 July 2025 (UTC)
July 2025 update
Happy summer, fellow editors in arms! I hope everyone is doing well; I have been less active, unfortunately, because of family issues, but I'm hoping to start my glorious return to unstoppable power editing with this update.
Headline: It has finally happened - we are below 50,000 articles! In the last two months, 14,859 articles were cleared. For yourself and your fellow editors, please clap.
Minutiae: For anyone interested in a more detailed breakdown of the numbers - average was 506.9 articles (thanks to juggernaut outlier December 2009); median 190. The stats of a month with a drive are completely skewed from a regular one so I won't bother. Needless to say, we smashed it.
Highlights: Almost all of 2009 is in dustbin of history! A number of non-consecutive categories have also met their inevitable end which is hopefully a trend that continues.
Low-hanging fruit: So many runty categories, dangling precariously like a sinner in the hands of an angry God, including June 2024 (48), January 2010 (49), and February 2023 (54).
High-hanging fruit: You know her, you love her — everyone's favourite BFC (Big Friendly Category), December 2009, is a scrumptious 5,388 articles as of this writing, after a decrease of 2,890, and is now the oldest category left standing. The other high-hanging fruit are, still, the Frustrating Five (name open for revision): January 2013 (750), May 2019 (1,345), June 2019 (2,940), September 2020 (917), and March 2024 (700). Once again, September 2020 had the lowest percentage of change between updates (2.29%). Godspeed to anyone working on these.
Challenge results: December 2015 beat January 2016, 194 to 201.
New challenge: The ties continue. August 2010 and September 2010 have weighed in at 141 and will fight to the death.
Announcements: Thank you all so much for the incredible work of the drive and the continued work reflected in the leaderboard. Now that the ball is moving so quickly towards zero(!), I think there's a question of how to find and identify unreferenced and untagged articles, which is a big concern of mine. Just yesterday I found references for Zeze ware, which was not tagged at all (sure, it had one external link but come on), so this is definitely an issue and I think one that we now have the time to focus on as the backlog becomes more manageable. In terms of project quality, I think these are the articles that will be the most detrimental to Wikipedia because they are flying completely under the radar. Are there any suggestions or ideas for how to identify these drifters? All the best, Kazamzam (talk) 23:40, 4 July 2025 (UTC)
@Kazamzam: Regarding that last point, we began identifying unreferenced and untagged BLPs during JUN25 (see here).
To summarize my comment there, at time of writing, we had 938 unsourced and untagged BLPs (as defined by the critieria on {{unreferenced}}) out of a total ~1.1 million. You can see the list of the 938 at User:ARandomName123/BLP. BLPs with no inline citations (ie. your article, but BLP) is around the mid 30 000s. Extrapolating this to the rest of the project (~7 mil articles) means around 6-7k unsourced articles (~245k with no inline). It's probably quite a bit more though, since normal articles are not held to as high a standard as BLPs, and there isn't an equivalent BLPPROD for unsourced articles. ARandomName123 (talk)Ping me!01:16, 5 July 2025 (UTC)
@Kazamzam And further to that, I have been slowly going through the list and tagging {{BLP unreferenced}} or otherwise dealing with articles. I aim for about ten a day, so as not to add a massive depressing lump of articles to the official backlog. Right now, we are at 676 articles still on the unsourced and untagged BLP list and about 30 on the backlog. So in a couple of months the list should be at zero.@ARandomName123, could we run your stricter script against the whole project (less Category:Living people I guess) to get an initial idea of possible unsourced and untagged non BLPs? Then we could begin whittling that down. Cheers, SunloungerFrog (talk) 07:29, 5 July 2025 (UTC)
Awesome @SunloungerFrog works great and I can now easily rank Unreferenced articles on specific topics by how popular they are. In order to prioritize (but also because they tend to be easier to find sources for). Cielquiparle (talk) 12:30, 5 July 2025 (UTC)
Doing... will take some time to go through all the articles, but currently (only a couple 100k in) the detection rate is ~1%. This is mostly due to all the [year] in [whatever] or [year] [event] articles, so hopefully it'll go down as it gets further. ARandomName123 (talk)Ping me!16:57, 6 July 2025 (UTC)
Roughly what proportion of unreferenced articles have no active projects?
Hopefully there is an easier way to estimate this than getting a complete list of the 625 active projects and pasting into Petscan “Has none of these templates”? Chidgk1 (talk) 08:43, 20 July 2025 (UTC)
I am not sure how you would get around active vs. inactive vs. semi-active WikiProjects. Of course, having an active WP does not mean that someone is actively looking at reports and prioritizing unsourced articles. I have noticed that there is a great deal of overlap between articles with no sources and articles that are stubs, articles with no short description, and sometimes articles without any WikiProjects listed. With the latter, there usually is not a TalkPage which might be a better approach for your query. However, this may take you to more of an AfD list, than a list of articles to improve. I guess it depends on why you think such a list would be useful. Rublamb (talk) 21:30, 20 July 2025 (UTC)
@Rublamb Thanks - so to simplify the question is it possible to estimate how many articles are:
Unreferenced
AND
over a year old
AND
have no talk page
The point of this would be that if there were a large number (say more than a thousand) to put in a proposal to have them all deleted, which would hopefully be less effort than you guys checking them individually. Chidgk1 (talk) 06:03, 23 July 2025 (UTC)
Of course such a list could be first notified to all active projects to give them time to add their project to articles before I or someone else puts in a bulk AfD for the reminder. Chidgk1 (talk) 07:04, 23 July 2025 (UTC)
@ARandomName123 Thank you for showing us your excellent code at User:ARandomName123/BLP.py. I like that it is short and in Python so more people have a chance to understand it and figure out whether it could do stuff like I want above.
Would you put this somewhere more official (not sure where) so that others can suggest changes and options? Also I would like to add more comments to the code as I don’t fully understand it yet. Chidgk1 (talk) 06:52, 23 July 2025 (UTC)
With an average of 10 BLPs per day being tagged as "Unreferenced" as part of our controlled release program, we have now accumulated a backlog of 92 or so Unreferenced BLPs. Please take a moment to look at [[Category: Unreferenced BLPs from July 2025]] and add some citations where you can. (Or deal with in another way after your best WP:BEFORE search.) Cielquiparle (talk) 06:19, 17 July 2025 (UTC)
@BeanieFan11 There seems to be a surprisingly large supply of notable NFL player bios which remain Unreferenced. They are generally very easy to find references for via Newspapers.com, etc. There were several of us during the June backlog drive who tried to stay on top of them, but it would be great if we could have support from the American football WikiProject as well. Cielquiparle (talk) 06:25, 17 July 2025 (UTC)
Progress update on the controlled release program - I've now gone through all of @ARandomName123's original search that returned 938 articles and tagged them appropriately. There are now some 399 unreferenced BLPs that are not already undergoing a deletion process. Let's get referencing! Cheers, SunloungerFrog (talk) 11:31, 9 September 2025 (UTC)
Nevertheless, in the end we had solid participation in July – all the more important since we are at a critical juncture given the discussions above which I won't recap here. Let's take a look at the July 2025 marathon leaderboard, shall we?
Hats off to the mighty Bearian who reached the top of the leaderboard early in the month and inspired the rest of us to hang in there and keep at it. (Incidentally, Bearian is now in 8th place for the year.) Thrilled also to see the return of Boleyn who powered through to a second place finish in July.
The next four users are in the 100+ club as well – JoeNMLC, Coldupnorth, JTtheOG, and yours truly. It was good to see some of our most competitive June backlog drive participants return in July with a top 10 finish – this includes JTtheOG and Bearian, but also Nayyn and FictionWitch. There is something to be said as well for editors such as JoeNMLC and ARandomName123 who show that steady participation over time matters (7th and 14th place, respectively for the year to date). Finally, welcome WikiOriginal-9 to the URA 2025 marathon!
Remember: It's easy to join in the fun in August. All you have to do is add at least one citation to articles tagged as Unreferenced (if needed), and remember to remove the Unreferenced tag (and replace it with another tag if needed). The overall 2025 marathon leaderboard is here and you will show up in the monthly column for August once you have removed Unreferenced tags on 5 articles. If you are diving into the trove of articles discussed above that are not yet tagged as Unreferenced, feel free to tag it as Unreferenced first, and then add a citation to a reliable source in a subsequent edit. (And equally...if you find mistagged articles that actually do have references, fixing those tags totally counts, unlike the monthly backlog drive rules.)
Many thanks to everyone who contributed in July and who are adding references in August. Let's keep at it. Happy referencing! Cielquiparle (talk) 01:09, 5 August 2025 (UTC)
@Arutoria, if you have dealt with fewer than five articles, it doesn't show on the leaderboard. But as soon as you've done five or higher, the correct number will appear. (I think that's right, anyway!) Cheers, SunloungerFrog (talk) 12:45, 19 August 2025 (UTC)
Thanks for reminding me. I've had this on my to-do list to fix for a while now, and finally got around to it today. It should now show your stats, including those less than five, as long as you were over five in any one month. @Arutoria's, for example, now contains the three from August, for a total of 16. The individual monthly leaderboards always shows all edits, regardless of count. ARandomName123 (talk)Ping me!03:42, 20 August 2025 (UTC)
Is there any way to fix the months counter of Unreferenced articles, so that it only counts the actual months?
I see that many editors are now manually tagging Talk pages as "Unreferenced" and obviously the months counter is now counting each of those Categories as months, driving up the number of months of Unreferenced articles up to 250 (up from 196).
@Cielquiparle manually tagging Talk pages....oy vey. I know that was a thing for articles for WP Hawaii using their project template that is theoretically useful but often doesn't get updated when references are added because it's not standard practice and/or most editors don't know about it. We might have to manually untag them, or perhaps our intrepid @DreamRimmer wants to take a crack at running a bot for this issue. I like the idea of tagging talk pages but in practice it's much more of a headache than it is a tool, in my experience. Also sorry to be away for so long but we rise. Thanks for the ping CQP! Kazamzam (talk) 16:05, 22 August 2025 (UTC)
Having investigated a little more, I think there are two problems.
Anything that isn't a monthly clean-up category relating to articles lacking sources needs to be removed from Category:Monthly clean-up category (Articles lacking sources) counter. Just to note that I tried this on the Albania category, but @Catfurball reverted the change, I'm not sure why. But maybe they can join the conversation
I have made many new unreferenced categories by country, since my new categories are not connected to templates if I remove the category from them will this take care of the problem. And later I could add it back when they get added to a template. @Cielquiparle:@SunloungerFrog:Catfurball (talk) 16:23, 22 August 2025 (UTC)
@Catfurball just don't add your unreffed articles by country categories to Category:Monthly clean-up category (Articles lacking sources) counter, and I think that should help fix the immediate problem. The wider question of whether it is desirable to add maintenance categories pertaining to articles to the corresponding talk pages (which sounds a bit bizarre and duplicative to me when one could run a petscan to get the same result) I leave to others wiser than me. Cheers, SunloungerFrog (talk) 16:40, 22 August 2025 (UTC)
Thanks @SunloungerFrog! (Yes the 185 looks correct as I always wondered why were counting those few random Talk page categories as their own "months".) I noticed some time ago that there were ~200 Talk pages marked as "unref=yes" for articles of interest to WikiProject Hawaii, but that pretty much all of them actually already had refs...and then someone finally went through and removed them all (possibly manually?). It is a big problem because no one can easily "see" that the Talk page is tagged as "unref=yes". Even worse, there seem to be some WikiProjects (specifically Cities), where "unref=yes" isn't recognized as a parameter, so all their tagged Talk pages end up appearing as their own separate categories. Cc: @Awkwafaba. Cielquiparle (talk) 22:29, 22 August 2025 (UTC)
@Cielquiparle: I was the one who removed almost all of the Hawaii articles, for almost all of them had references. And yesterday I did the same thing with Michigan, there were 42 marked as being unreferenced, when actually there was only 4. I have doing this cleanup since last year of the talk pages. Catfurball (talk) 15:29, 23 August 2025 (UTC)
Hi All, there seem to be quite a few unreferenced music albums. I sometimes struggle with references for these even though clearly the albums exist as they are on a variety of platforms to listen to. Sometimes there are news articles although its harder for older or lesser known albums. I am wondering what experience other users have in references for these? In most instances, Discogs always seems to have an unreferenced album but Talk:Discogs seems to say its not sufficient to be a reference as its user generated content, which I get but then isn't Wikipedia... ? Anyway I was just wondering what other users do for albums here? Thanks Coldupnorth (talk) 21:46, 24 August 2025 (UTC)
I've primarily been using Newspapers.com while going through the albums backlog. Though that only helps with the English ones. There's a ton of German, Swedish, and Finnish death metal albums that I have no idea how to reference. I doubt the metal specific websites would be covering those either, even with them being on the subject genre. SilverserenC22:56, 26 August 2025 (UTC)
Future backlog drives schedule
The next big Unreferenced articles backlog drive is in November 2025 (per previous discussions).
For 2026, I have presumptively grabbed March 2026, August 2026, and December 2026 for now on the WP:DRIVES annual calendar. (At first I grabbed April 2026 which remains completely empty but remembered ARandomName123 might have exams and a lot of people tend to disappear in April anyway.)
But please let me know if these timings don't work and we should adjust. I mean, I know it's far in the future and it's hard to predict that far out, but it's a bit competitive, so... Cc: @ARandomName123, @SunloungerFrog, @Kazamzam. And if 3 is too much in a year, we could always let go of one later. Cielquiparle (talk) 04:10, 27 August 2025 (UTC)
Thanks for considering my exams! March and August 2026 work for me. I have fall-term exams in the first two-thirds of December, but we're still over a year out from that so we can deal with it when we get closer. ARandomName123 (talk)Ping me!00:39, 28 August 2025 (UTC)
Unsourced and untagged articles
After Kazamzam brought it up in the July update, I ran a program over all articles on Wikipedia which tried to identify all unsourced articles, that were still untagged. Any of the following were counted as a reference:
Anything within Category:External link templates and its subcategories (minus {{Authority control}} and {{Authority control (arts)}}). This contained a lot of stuff that can't really be counted as references, but I gave up trying to manually sort through useful and not useful ones after like an hour or so.
Any line containing "==External links==", "References==" or "sources==" followed by a *. This is for those bulleted references that don't use any specific reference templates
Any URLs
The standard reference things, like <ref>, {{sfn}}, {{harvnb}}, {{cite}}, {{refn}}.
Any template ending in "index}}", or starting with "{{list of lists" for those list articles
Some other misc. templates I came across (like {{surname}})
The results are at User:ARandomName123/unref1, 2, 3, 4, 5, 6, 7, 8. (split up so it was easier for me to copy-paste them in). The total count is 65957. There are quite a bit more templates that contain sources, or can be used to exclude articles, so if you do come across one in my lists, please let me know and I'll filter them out across all 8. A spot-check shows a decent amount of the list still contain some sort of reference through some random template, so more filtering would definitely be helpful. Any other suggestions/improvements are also welcome. Cheers, ARandomName123 (talk)Ping me!02:15, 16 July 2025 (UTC)
Seems to be finding a few which have already been tagged “Unsourced” and “Unref” and "No sources" - I don’t know how to find what else redirects to “Unreferenced”
Some stuff I found in curly brackets “Larissa div” "Polygyros div" (therefore I suspect some other Greek divs might be similar) “US-airport-minor” “MinorPlanetNameMeaningsFooter” "Animal common name" "Plant common name" (maybe are others like fungi)
I am not a member of any of the Category:Health WikiProjects but I think it would be really useful if you could run this for real to tag all health articles which might be uncited as uncited. I am sure it is better for some of these to be mistakenly tagged as uncited than for uncited medical articles to lurk undetected.
Thanks, the stuff you mentioned has been added, and I'm currently running it again. Imdb would probably be counted. No footnotes would also be counted. Bibliography probably not, since they are sometimes just listings of the authors works. Michael Budde and Methoxymethylenetriphenylphosphorane: added {{Short pages monitor}} and {{ref}} to the exclusion criteria. Red links were deleted between the run through, and me posting it here. I will look into the other templates. ARandomName123 (talk)Ping me!17:01, 16 July 2025 (UTC)
Regarding "No footnotes", does this not sweep in a lot of articles which are not actually unreferenced but just lacking inline citations? (a common feature of translations from de-Wiki, & doubtless others). Ingratis (talk) 17:55, 16 July 2025 (UTC)
My main concern was that "No footnotes" might be incorrectly applied. You can see which of the listed articles have that template through petscan:37243882, though I'm still culling the lists with some other criteria. I'll check it once it's completed, and if there's a large false positive rate, I'll add it to the exclusion criteria. ARandomName123 (talk)Ping me!18:45, 16 July 2025 (UTC)
Do you have a specific wikiproject in mind for the health one? Not sure how I can get Petscan to work with that (maybe @SunloungerFrog can? you're pretty experienced with Petscan)
Lists that are primarily navigational in nature do not traditionally contain citations. As an example, I see the List of Turkish physicians in the Petscan results for Turkey, which contains nothing except an alphabetized list of links.
Thanks, @Chidgk1. See below for an evaluation of the false positives. If you'd like, I can also provide a list of non-list pages in the Turkey list. As for running it for real, we could just manually tag them, since a bot would take a while to get approved.
Warrior relief of Efrenk: Same as above, but I'm not sure if I want to include Bibliography as an exclusion criteria, it could also refer to an author's list of works.
Lycus (river of Cilicia): Added {{SmithDGRG}} to the exclusion criteria. The template it redirects to should already be excluded, but it wasn't caught because it used a redirect.
@ARandomName123 Thank you but I am against manual tagging because WP:Turkey was just an unimportant test. Whereas the medical stuff could have bad effects in the real world. So given the very small proportion of wrongly tagged articles on WP:Turkey I strongly support running your excellent creation to really tag anything on the whole of Wikipedia we suspect might be uncited. I contend that the benefits of flushing out unsourced medical articles will greatly outweigh the harm of wrongly tagging some cited articles as uncited. Chidgk1 (talk) 07:01, 19 July 2025 (UTC)
Feel free to remove the links as you see fit, but I've also been updating it whenever a new template/criteria to exclude is discovered. ARandomName123 (talk)Ping me!21:21, 19 July 2025 (UTC)
I'm looking at the health-related Petscan list. 60% are lists that I would not expect to be sourced.
Nitrosoproline is an edge case; technically, there are external links on the page (in the {{chembox}}; in this case, the PubChem link verifies half of the first sentence) but I'd expect ordinary citations as well. The same is true for Hydrocortisone phosphate, which has {{drugbox}}.
Alveolus looks almost like a disambiguation page. (BD2412, would you like to take a look?)
@ARandomName123 Thanks for all your hard work. Are you able to remove lists in the original code? I think this should be run as a bot so that everyone can see what is uncited not just people who end up here. I have never done a bot before but if you don’t want to go through the bureaucracy could you let us see the code so I can try to get it approved? I have some (extremely out of date) coding experience so may be able to tweak it. Chidgk1 (talk) 08:10, 20 July 2025 (UTC)
@Chidgk1: I can remove the lists if necessary. GreenC below mentioned that they already went through the approval process back in 2019, so we could just try using their bot to skip the approval process. ARandomName123 (talk)Ping me!17:14, 20 July 2025 (UTC)
Pardon me if it's referenced above and I missed it, but GreenC did a bot run with this purpose in 2019. I believe all the relevant info is linked from this VPP thread. GreenC purposefully took a relatively conservative approach to tagging, but you may find his notes helpful. The final BRFA is here. Ajpolino (talk) 11:21, 20 July 2025 (UTC)
Thanks for the ping. Yes I did make a program that detects articles with no refs. It is conservative and went through extensive manual testing with multiple editors helping. How many it tags is a fuzzy recipe of what constitutes an "article" and what constitutes a "reference". It managed to tag about 10,000 articles in 2019, and AFAIK nobody complained of false positives. It worked. Getting RfC approval was very difficult, many editors were strongly against, fearing it would go haywire. Between resistance from the community, difficulty of the code with endless edge cases, the difficulty of testing, I vowed to give this project a rest! However if there is a request to run again now, 6 years later, I probably still have permission. The first run proved it worked. -- GreenC16:27, 20 July 2025 (UTC)
@GreenC: Oh, that's great! Would it be possible to do a non-editing run of your bot to get an idea of how many pages match your criteria? If we were to run it again, I think we would prefer it to be trickled in rather than in one big dump (see below for what we're doing for BLPs). ARandomName123 (talk)Ping me!17:18, 20 July 2025 (UTC)
ARandomName123: Checking the code, apparently I made some options (memory is hazy):
# 0 = post results to test page
# 1 = live run and post test results
# 2 = live run only
# 3 = save list of articles to filename G["static"] "list"
# 4 = live run and save list of articles to filename G["static"] "list"
# Remove template
# 5 = save list of articles to filename G["static"] "list-opposite"
# If 0, continue to end
# Otherwise number of hits to discover before stop processing
MaxCount = 0
Looks like out of the box it could do what you want. Do you prefer #0 or #3? The later would be a text file. There is also the intriguing #5, which essentially is the opposite: pages with the template that may not need the template anymore (according to the bot's criteria). As for MaxCount, it would stop after a certain number of edits. When it runs again, it picks up at the same spot, so it doesn't process the same articles early in the list. This way it could be setup to run from cron automatically once a month, each month gets a maximum number of edits then it stops for that run. -- GreenC18:42, 20 July 2025 (UTC)
OK. It downloads 7+ million articles and does various regexing routines. It will likely take weeks to finish, linear not multithreaded. How about 500 results and check it out before doing the whole enwiki. To make sure it's still working correctly after 6 years. I'll post when the 500 are done so you can check it out. -- GreenC19:15, 20 July 2025 (UTC)
ARandomName123: User:GreenC/data/noref/July2025restart .. It found 16 in 167,000 pages before I stopped it. At this rate (the article feed list is randomly sorted), it will be a total of about 700 in 7 million pages. Assuming these 16 look OK - I have not looked yet - suggest letting the bot run through to completion in live mode. -- GreenC16:38, 21 July 2025 (UTC)
@GreenC Thank you very much. Running your “conservative” bot all the way through again live is a great first step I think. After that it would be very interesting to run @ARandomName123s bot in dummy mode and see how many it still finds in total and check a sample manually to estimate what proportion it then thinks are unreferenced but are actually referenced.
Also we could maybe compare the criteria of the 2 programs and see how they are different.
The long runtime does not matter I think. Is the environmental impact such as carbon footprint significant? Seems unlikely it would be nowadays, but if it is then given your “green” user name is it possible to tell the bots to run at times to minimize their environmental impact? Chidgk1 (talk) 06:20, 23 July 2025 (UTC)
@Chidgk1: I believe the environmental impact of us running these programs is pretty much negligible. The thing about the long runtime is that I run it locally on my computer, so I need to leave it on for it to run. I should probably move it to toolforge or something.
Regarding criteria, I mentioned mine above (have since added Category:Specific source templates, any alphabetical character below a reference header, outlines, and some other misc. templates). GreenC's is at User:GreenC bot/Job 11/How. Mine is far more liberal with tagging, and as such, is more likely to find false positives. I believe our template filtering is somewhat similar, he says he has 6k+, and mine is currently a bit more than 6185. I believe one of the main differences is that his bot also skips stubs and matching section headers. ARandomName123 (talk)Ping me!02:25, 24 July 2025 (UTC)
Great, thanks. Running through to completion in live mode sounds fine to me, just try not to have it dump all 700 tags at once. ARandomName123 (talk)Ping me!21:16, 21 July 2025 (UTC)
Rublamb, I ran it on 200 articles that have the template, and it found 200 hits. I think the reason is my program is so conservative in determining when to add the template ie. the opposite of extreme conservativism is extreme liberality. -- GreenC17:09, 21 July 2025 (UTC)
@Chidgk1 GreenC is saying that the bot thinks nearly all articles tagged as unreferenced actually have references, because its definition of a "reference" is very broad. Toadspike[Talk]10:11, 23 July 2025 (UTC)
Yes, thanks for the succinct explanation. It's kind of a brain twister when you start thinking about it. Ultimately the considerations for adding the template, versus removing it, are different. When it's a bot doing the adding and removing. — GreenC14:41, 23 July 2025 (UTC)
Following the chat about a putatative Nov 2025 drive, I thought I would do a bit of petscan analysis of the likely unreferenced blob. So, as of 30 September 2025:
Out of 65957 articles originally found by ARandomName123's script
I ran the script through it again, and the current count is 39817 articles. The 65957 total was the original, and a lot of it was removed when more templates were discovered above. ARandomName123 (talk)Ping me!03:17, 1 October 2025 (UTC)
Thanks @ARandomName123. Maybe it really was just the handful that got converted to redirects in the last couple months. Anyway another thing I've noticed exactly twice is articles already tagged as Unreferenced, but which use {{noref}} instead of any of the other variations. (Pretty sure I manually changed that both times. Noisemaker was one of them.) Cielquiparle (talk) 09:08, 9 October 2025 (UTC)
@Cielquiparle: For the redirects, I just forgot to add a filter for them (whoops lol). The original search through all articles had a built-in filter, but the search on the list didn't and I forgot to add one.
As for noref, I've added it to the filter. Please let me know if you come across others, I don't have all the template redirects for all the templates. ARandomName123 (talk)Ping me!15:19, 9 October 2025 (UTC)
@ARandomName123 On the subject of legitimate {{No footnotes}} or {{Inline citations}} tags continuing to sit in the "likely unreferenced" blob...could we try filtering out any articles that have either of those tags *and* have at least one source listed under the heading "Further reading" with a bulletpoint? If so, we could make sure that any articles that are confirmed to be correctly referenced as {{No footnotes}} or {{Inline citations}} have at least one source listed under "Further reading". (Not all of them do but we could make it true.) For now I wanted to see what happens if the required heading is "Further reading" as other headings like "Bibliography" often turn out to be a list of works written by the subject, and that requires some other tag or two. Cielquiparle (talk) 06:36, 15 October 2025 (UTC)
Hats off to everyone who added citations to Unreferenced articles in August 2025! It was a tough month, as we battled the flood of newly tagged articles...but we somehow made it to the other side.
In August, we had a whopping 8 editors who logged 100+ points, with 2 editors ahead of the pack at 200+. They were followed by 7 more editors who each had 31+ points, the equivalent of 1 a day.
The race for the top spot seemed rather intense, as @Bearian appeared to be keeping the crown for the second month in a row, but was ultimately surpassed by @Silver seren. Congratulations to both.
Speaking of people who don't like to lose: It was great to see serial backlog drive champion @JTtheOG back in action to claim bronze for the month – and incidentally, for the marathon year-to-date. (And yes, I noticed this especially because there was a fleeting period when I managed to climb to third place myself, only to get knocked back down.)
And finally, a shout out to @Miraclepine and all other editors who joined in on the action for the first time last month. Welcome! English Wikipedia still has 47,000+ articles tagged as Unreferenced (and many more that have yet to be tagged). Sorting through these articles, trying to verify facts, and adding citations to reliable sources wherever possible is important work, and we need all the help we can get.
It's still important work, @Miraclepine. Thanks for fixing all those tags. (And also...the ambiguity between DAB pages vs. set indices vs. lists is challenging to deal with. I am also coming across pages which don't seem to it well in any of those categories which seem to be padding the overall count.) Cielquiparle (talk) 09:23, 3 September 2025 (UTC)
You were really in a battle for your life there at the end, Cielquiparle. That four way fight was kind of intense for a bit. SilverserenC04:13, 3 September 2025 (UTC)
Thanks. It's only going to get harder from here on out though. I'm running out of straightforward lists to go through. We're eventually just going to have only an eclectic mix of article subjects that are annoying to reference. SilverserenC05:14, 3 September 2025 (UTC)
Species articles are pretty easy to reference, especially if they're really short. POWO works as a good reference, and some genera have papers online with a description of every species in that genus, and so can be used as a reference for those species articles. ARandomName123 (talk)Ping me!06:10, 30 September 2025 (UTC)
Excellent, @SunloungerFrog. I have been wondering and wanting exactly this. (Relieved the total number is below 90k. Though...did the total number really go down by 1,292 in one day?) Cielquiparle (talk) 03:21, 2 October 2025 (UTC)
What ARandomName123 said, rather more lucidly than I would have. I'll note that the way I do the counting is not exact exact, because (for example)
if editor Alice looked at probable unreferenced article Foo on Monday and tagged it properly with {{Unreferenced}}
then editor Bob added a reference and removed the tag on Tuesday
it would still appear in the stats for probable unreferenced articles on Wednesday, because it wouldn't necessarily have been removed from ARandomName123's source lists, and doesn't contain {{Unreferenced}}.
The article would only be removed entirely from our stats when ARandomName123's source lists are refreshed, and the entry removed from there. The good thing is that the totals I report will always be more pessimistic than the actuality, so no one will have a false sense of achievement, but rather be nicely surprised when the source lists are trued up, so to speak. Cheers, SunloungerFrog (talk) 08:14, 2 October 2025 (UTC)
@SunloungerFrog: Nice job! Love the colour coding and subdivisions, and I'm also relieved and surprised the total is below 100k, let alone 90k. It would be nice to add to the "Progress" section. Automated updating would be best, would you like to set that up? Alternatively, I can write and/or run it on my bot if needed.
@ARandomName123: If you are able to write and run it on your bot, that would be quickest - thank you! It'll take me rather longer.... Let me know if I can help - hopefully the bash script is somewhat useful. Just further to my point above, it would be helpful if you were to run your script to refresh the source lists before running the stats. I don't know how long that refresh takes, so maybe it isn't feasible to do that every day. Maybe a weekly refresh and then the stats? Cheers, SunloungerFrog (talk) 08:19, 2 October 2025 (UTC)
@SunloungerFrog: It should still be feasible to do it daily as the refresh should only take one to two hours. It only took much longer last time because I was running it locally in the background of my laptop during class, which I shut off in between classes and when I headed home. I'll make it update like 4 hours before the chart to be safe, or run the chart immediately after the refresh. ARandomName123 (talk)Ping me!05:37, 3 October 2025 (UTC)
@SunloungerFrog: Up and running now. The unref pages (1-8) are updated starting at 22:00 UTC, taking roughly 60-80 minutes to run. The table, which I have copied to User:ARandomBot321/URAstats, updates at 00:00 UTC, meaning the number for a date is the number at the very start of that date. I can adjust that if you'd like. ARandomName123 (talk)Ping me!00:42, 6 October 2025 (UTC)
Amazing work - thank you! The timings look fine to me. I've now included the chart in the progress box. Feel free to revert if you think it needs more discussion - I know the colours are a bit garish.@ARandomName123: I know you will have thought of this already, but we will have to box clever as the chart fills up, and at some point institute a "one in, one out" mechanism to age off the oldest bar as the new one is added so that it doesn't get too crowded. Cheers, SunloungerFrog (talk) 18:20, 6 October 2025 (UTC)
@Cielquiparle Well. I sort of have a solution to that, in that I have ported the chart to use the new MediaWiki Chart extension. It is a bit more elegant, in that
the data and the chart definition are separated, so that you can use the same data in a bunch of charts
the data can be viewed in tabular format, which is quite nice
the charts themselves are a bit more interactive - one can turn on/off different lines of data, hovering over them shows that day's numbers, etc.
the charts deal with sizing rather more gracefully, so we don't have to worry about the "one in, one out" thing.
The trickier thing is that the bot would need to be configured to update the data table on Commons here c:Data:URAstats Data.tab, and I don't know how much of a pain that is, whether there are bot approvals needed etc. etc. - any thoughts, @ARandomName123? As an interim measure, you could update a page on enwiki and then I could cut'n'paste it across on a daily basis?
@SunloungerFrog: Oh wow, I love that last one (area chart). Also has the advantage of not clumping together bars and showing the total. Thanks for setting those up! I'll look into the bot process on Commons. As for the interim measure, do you want it formatted the same as it would be on Commons? ARandomName123 (talk)Ping me!15:16, 16 October 2025 (UTC)
@ARandomName123, the area chart is my favourite too; my only gripe is that the y axis doesn't start at zero. For the interim measure, if you can format the data as it's needed on Commons, i.e. a JSON array (I think!), then I am happy to just cut and paste it across as and when. Thank you! Cheers, SunloungerFrog (talk) 15:36, 16 October 2025 (UTC)
@SunloungerFrog Since the area chart doesn't do zero, how about flipping or re-ordering the stack, such that the bottom category that gets cut off is "likely unreferenced" and the main visual focus is "Unreferenced articles"? Or is there no way to do that? Cielquiparle (talk) 02:50, 17 October 2025 (UTC)
So did some digging around, and came across phab:T385344, which SunloungerFrog seems to have already found. No changes in the last few months though. They did have a suggestion of writing a Lua module using mw:Extension:Chart/Transforms, but I got no clue how that works.
However, I was able to get it to include zero by switching around the ordering of the data. I did some test edits to the data pages, which can be seen in the history of the data tab. Turns out, if you put the smallest one at the bottom (not sure about the others, I only tested the smallest, which was BLPs), you can get the graph to start at zero!
Also turns out the area chart is really cool for another reason. If you click on the legend at the top, you can remove stats from the view! Which means you can view only the change in unreferenced articles, or only the change in unreferenced BLPs or even a combination of two or more. ARandomName123 (talk)Ping me!04:32, 17 October 2025 (UTC)
Yes, the ability to click on/off of individual stats is excellent, isn't it? And nice hack on putting the unreferenced BLPs first to get y=0. I have now made that change permanently, and added in the data for today, the 17th. @ARandomName123 if you are able to edit your bot so as to switch the order of unreferenced BLPs and unreferenced articles in the JSON, so that BLPs are first, then articles, that would be very helpful. I've put the new area chart live in the progress box. Cheers, SunloungerFrog (talk) 06:09, 17 October 2025 (UTC)
@Cielquiparle: The big decrease is mostly from the yellow group (Probably unreferenced articles), which decreased by 1149. I ran an update over the pages that section pulled from, which decreased the amount by roughly a thousand compared to Progress update by SunloungerFrog, which roughly lines up with 1149 (see my comment above, example update to unref1: Special:diff/1314329587) The update ran from ~18:00 to 21:00 UTC on the 30th, so it was probably between the times when the data was grabbed on the 30th and 1st. ARandomName123 (talk)Ping me!05:19, 2 October 2025 (UTC)
Still not displaying properly for me. Lozenge is half the height of the number which gets cut off. In dark mode the number goes grey. Cielquiparle (talk) 17:33, 20 October 2025 (UTC)
Thank you both. WhatamIdoing (talk) 20:19, 7 March 2025 (UTC)
Sent – DreamRimmer ■ 16:16, 21 May 2025 (UTC)
Doing... will take some time to go through all the articles, but currently (only a couple 100k in) the detection rate is ~1%. This is mostly due to all the [year] in [whatever] or [year] [event] articles, so hopefully it'll go down as it gets further. ARandomName123 (talk)Ping me! 16:57, 6 July 2025 (UTC)
