Wikipedia talk:WikiProject YouTube

Upcoming transition to automated updating of subscriber counts

Hi all! Following up on the consensus here, @BrokenSegue and I are preparing to implement a system that will update the subscriber counts in {{Infobox YouTube personality}} automatically, using a module that sources the data from BrokenSegue's bot on Wikidata. This should both help these articles stay more up to date and also reduce watchlist clutter (as the data updates will all happen at Wikidata). Please feel free to let us know if you have any questions or concerns. I'll post again once the system goes live, at which point please let us know of any issues you encounter. Cheers, {{u|Sdkb}} ^talk 18:31, 28 February 2023 (UTC)

I don't necessarily like the idea of the counts being hosted at WikiData, since putting them in a submodule would work just as well, but I'm glad to see this is finally getting off the ground; I know there are a few folks that have been poking around the edges of this for a few years now. Primefac (talk) 12:18, 1 March 2023 (UTC)

I don't think this is a good idea. Many articles display

"[...] subscribers (main channel)

[...] subscribers (combined)", with an Efn that contains the subscribers counts of all channels owned by the article subject.

This change completely breaks this. Strugglehouse (talk) 13:21, 25 March 2023 (UTC)

I think it would be a good idea to create a template to wrap the module, and modifying the module to allow for specifying which channel to grab stats from. This would allow for doing the combined subscribers and view counts. Or add a new function to the module that would output and format with multiple channels listed in Wikidata. (Side note: I'm interested in migrating away from |channel_name=, |channel_url=, |channel_direct_url=, and |channel_display_name= and into {{YouTube channel}}, which dovetails nicely from this issue raised.) SWinxy (talk) 17:13, 25 March 2023 (UTC)

@Strugglehouse, thanks for bringing this up! Could you provide a link or two to examples? This seems like a fairly rare use case, but it is something we'll want to find a way to handle. Cheers, {{u|Sdkb}} ^talk 22:47, 25 March 2023 (UTC)

Not uncommon, surprisingly. Dream (YouTuber), MrBeast DanTDM, and Behzinga, among others. (Nowadays creating multiple channels is less of a thing; I don't expect YouTubers from the last 4 years to have second channels.) SWinxy (talk) 23:53, 25 March 2023 (UTC)

If you look at many YouTuber's Wikipedia page, you can see examples of this. If you see the ones that Swinxy has linked, as well as others such as Jaiden Animations and Wilbur Soot, you can see examples of this formatting under the "Views" section in their Infobox, and you can see that the "Subscribers" section no longer works like this. Strugglehouse (talk) 10:37, 26 March 2023 (UTC)

Early-stage launch

Update: We've now turned on the automated subscriber counts for most articles. Any that produce errors will fall back on the manual parameters and be placed in a tracking category. Once we've eliminated the errors, we'll be able to deprecate the manual parameters to complete the work. Please let @BrokenSegue and I know if you see anything amiss. Cheers, {{u|Sdkb}} ^talk 04:29, 23 March 2023 (UTC)

Can the template segregate mainspace articles from draftspace and userspace ones? e.g. [[Category:Pages with YouTubeSubscribers module errors| ]] SWinxy (talk) 04:49, 23 March 2023 (UTC)

@SWinxy, we could wrap the tracking category in {{ns0}} to get rid of stuff outside mainspace. Would that be helpful or is there a better approach? Cheers, {{u|Sdkb}} ^talk 04:55, 23 March 2023 (UTC)

Yeah I'd like that. I wasn't sure if I should suggest that either. SWinxy (talk) 14:48, 23 March 2023 (UTC)

Sorry, i'm not sure I understand what's being suggested. You want me to wrap the category statement in that template when it's not in main namespace? BrokenSegue 17:12, 23 March 2023 (UTC)

@SWinxy @BrokenSegue, I implemented the fix, so pages in namespaces outside of mainspace are no longer being added to the error tracking category. It looks like there are about 850 entries in the tracking category out of 2500 total transclusions of the infobox. Doing some spot checks, it appears that the culprit in almost all cases is that the subject's Wikidata entry doesn't have any information about their YouTube channel. Fixing that would be a good gnoming task, or maybe it's possible to bulk import the channel parameter data to Wikidata somehow. {{u|Sdkb}} ^talk 17:17, 23 March 2023 (UTC)

Another source of errors are pages that have multiple YT channels associated but no one is marked as preferred. The bot only populates the subscriber count when there is a single "primary" channel. BrokenSegue 17:28, 23 March 2023 (UTC)

Wow, this sounds great. Thanks! InfiniteNexus (talk) 16:18, 23 March 2023 (UTC)

I am confused as to how this is supposed to work. Does it display the subscriber count and date, overriding whatever explicit values are set in the infobox? If so, does that mean we should remove the explicit |subscribers= and |subscriber_date= parameters from each article so as not to confuse editors that setting them has any effect? If an article shows up at Category:Pages with YouTubeSubscribers module errors, how do we tell what the error is that causes the article to have been placed there? Maybe the sorting for that category ought to be by the error, similar to how infoboxes with a category for articles using an unknown parameter sort them by the unknown parameter name instead of by the article name.  — Archer1234 (t·c) 09:30, 24 March 2023 (UTC)

It looks to override them, yes. Each parameter independently falls back to the user-set values when it encounters an error. As Sdkb says, they look to be a result of missing data on Wikidata. SWinxy (talk) 21:25, 24 March 2023 (UTC)

Yes, that's correct. We'll eventually want to remove the manually entered data once the parameters have been deprecated; I believe that can be done by PearBOT or something similar. But no need to yet unless there's an article you particularly care about.

Regarding how to tell what the error is, if you use {{#invoke:YouTubeSubscribers|subCountNice}} on an article and preview it, it may tell you. You'll want to go to the article's associated Wikidata item (Tools > Wikidata item) to investigate directly. Cheers, {{u|Sdkb}} ^talk 22:55, 24 March 2023 (UTC)

Sdkb, perhaps the error could always be shown in preview and in bold red? (when {{REVISIONID}} is empty) — Alexis Jazz (talk or ping me) 20:34, 30 May 2023 (UTC)

The infobox of Destiny (streamer) is glitched

Hi there, As brought up on Talk:Destiny (streamer)#Updating YouTube info, and confirmed by myself in the edit history of Destiny (streamer), the article's infobox is completely glitched. Altering the "subscribers" and "views" data points doesn't result in any changes; It keeps displaying 406,000 subs no matter what.

If anyone knows what might be causing this and/or knows a solution to fix this, it'd be appreciated. CeltBrowne (talk) 21:18, 27 March 2023 (UTC)

I take it that it's related to the switch to automated subscriber counts discussed in the above section. CeltBrowne (talk) 21:24, 27 March 2023 (UTC)

The template overrides inputted numbers if it can successfully grab things from Wikidata. But in this case, the bot isn't updating the channel statistics. I think because of problems like this and above that the current template code should be changed so that it will only run as a backup or invoked manually. I've updated his wikidata page and it should propagate shortly. nvm it was because the module takes from P8687 instead of P2397. It was an instant update. SWinxy (talk) 22:31, 27 March 2023 (UTC)

@SWinxy: the bot (which I operate) only runs once a week. is that not frequent enough? BrokenSegue 05:29, 28 March 2023 (UTC)

The bot's not updating his statistics. See wikidata:Q60320906#P8687 and wikidata:Q60320906#P2397, as besides the main channel which I manually updated yesterday, it's all out of date. SWinxy (talk) 15:39, 28 March 2023 (UTC)

Is there a reason why this isn't on enWiki to allow for easier updating if things are out-of-sync? Primefac (talk) 18:09, 28 March 2023 (UTC) (please do not ping on reply)

Why would it be easier to have a bot update the numbers on enwiki instead of wikidata? That would mean maintaining two bots (one for enwiki and one for wikidata) for no reason. BrokenSegue 19:42, 28 March 2023 (UTC)

besides the main channel data that you added under P8687 there is no youtube data whatsoever (it's all twitter data). This is because until I updated the P2397 to point at a singular best channel the lua module wasn't sourcing data from wikidata. The way the bot works (which could be changed if there's consensus) is that it looks for a singular best main channel and tracks the subscriber numbers for that channel. Destiny had no marked best channel and so no data was being tracked. BrokenSegue 19:41, 28 March 2023 (UTC)

That behavior is kinda weird tho. Today I had someone in the Discord ask why their edits to the subscriber field wasn't showing on The 2 Johnnies. The behavior of the bot and the behavior of the module differ in that the bot's update scope is narrower than what the module pulls. Where would consensus need to be established for the altering of the bot? SWinxy (talk) 20:33, 28 March 2023 (UTC)

The bot and the module are in agreement I think (I wrote both of them). Each item has one main youtube channel and that is the subscriber count that we track over time. If there isn't a main youtube channel then we populate nothing and grab nothing. I think we can establish consensus for changes here as long as whatever we change is within the remit that was approved on Wikidata (which is pretty broad). BrokenSegue 22:54, 28 March 2023 (UTC)

The 2 Johnnies has this problem. The module is grabbing the data from Wikidata and putting it in the infobox. But those numbers are stale. SWinxy (talk) 23:06, 28 March 2023 (UTC)

Ah ok I understand the issue for that page. The bot follows the following rules on when to update the count: 1) has the number of subs grown by 10%? 2) is the current sub count over 1 year old? In the case of the page you reference the data hasn't changed 10% and is less than 1 year old. These parameters could be tuned. Or else a human could manually add a new subcount to wikidata. What if I made a webpage that allowed anyone to click a button and summon the bot to update the count immediately (maybe with some sane limits like > 1 week and >2% subs)? I agree that the internal functioning of this process is not transparent but I'm not clear how it could be made better. BrokenSegue 02:00, 29 March 2023 (UTC)

What's next

The change has led to confusion among editors unaware that their additions to infoboxes are being overridden by the module mixed with not great automation behavior. Here's a proposal on what changes I think need to be made:

1. Change the code in the infobox to not override manual entries, falling back to the module: {{#if: {{{subscribers|}}} | {{{subscribers}}} | {{#invoke:YouTubeSubscribers|subCountNice}}}}
2. Update the module to pull all entries in P2397 unless there are preferred entries (in which case only those would be pulled)
3. Change the bot's behavior to query for all entries in P2397, not just the preferred one
4. Deprecate P8687 in favor of P2397 (Wikidata:Project_chat/Archive/2022/03#social_media_followers_(P8687)_was_a_terrible_idea|discussion)

SWinxy (talk) 20:53, 28 March 2023 (UTC)

I'm indifferent on 1. Against 4 (I think it's a non-starter). And 2 and 3 are both options but I don't think it's clear what it means for the module to pull all entries. What number will it return? The maximum? The most recent? BrokenSegue 22:51, 28 March 2023 (UTC)

I should have specified the output of the module for #2. Format it the way that many articles have: {{ubl|123,000 (channel A)|56,000 (channel B)|10,500 (channel C)}} etc. List 3 channels by default, and anything more would have the third entry be "combined", e.g. MrBeast. SWinxy (talk) 23:12, 28 March 2023 (UTC)

1 has to be implemented. There are several articles where the Subscribers paramter is used to display subscribers for multiple channels, e.g. over at Linus Sebastian and Rhett & Link, where both creators have numerous YouTube channels. Having automated subscriber counts means only one subscriber number will be displayed, which doesn't accurately reflect either creator's total subscriber base. - Evelyn Marie (leave a message · contributions) 08:27, 1 April 2023 (UTC)

I'm unclear what "total subscriber base" even means. You can't just add the number of subscribers across all their channels. There is considerable overlap between those? BrokenSegue 15:51, 1 April 2023 (UTC)

@BrokenSegue: total subscriber base = total number of subscribers across all channels. and yes, you literally can add up the total number of subscribers. have you checked either article? i'm referring to the "main" and "combined" numbers several articles about YouTube creators use. - Evelyn Marie (leave a message · contributions) 21:26, 1 April 2023 (UTC)

Yes I did check either article. And yes I know one can just add up the numbers. I'm questioning if that number is meaningful. BrokenSegue 00:29, 2 April 2023 (UTC)

It is. Various creators spread their content across multiple channels. I think its useful to have the combined total of all subscribers, especially if the infobox has a list of all YouTube channels. Having one figure is NOT representative of a creator's total following or fanbase. - Evelyn Marie (leave a message · contributions) 03:25, 2 April 2023 (UTC)

Sure but a youtuber with 2 1million sub channels is not comparable to a youtuber with 1 2 million sub channels. It seems odd. Are there examples of non-Wikipedia sources using this metric? BrokenSegue 04:35, 2 April 2023 (UTC)

No, but why does it matter? The primary channel shouldn't be the only channel that counts towards a creator's total subscriber count. - Evelyn Marie (leave a message · contributions) 05:49, 2 April 2023 (UTC)

I share BrokenSegue's view that adding up subscriber counts is a misuse of parameter. If two channels are closely related enough that someone seeks to have the presented together in the same infobox, they're also presumably going to have significant subscriber overlap, so adding them up would count many subscribers multiple times, artificially inflating the count.

To answer the direct question above, the degree to which reliable sources use the metric matters, because Wikipedia is supposed to be based on such sources.

I think that the proper way to talk about subscribers for such channels is a question that goes beyond the direct scope of what we're doing here. I'd suggest opening a separate level-2 thread on that to try to reach a consensus among project participants. Once you have that, feel free to let us know and we can figure out a way to implement it. Cheers, {{u|Sdkb}} ^talk 21:56, 2 April 2023 (UTC)

Linus Sebastian for example, has multiple channels under his Linus Media Group banner. I think its very wrong to only include the primary channel. It cancels out people who may subscribe to one channel but not the other. That's why I honestly severely disagree with even implementing Wikidata support in this instance. Or, there should be at least a way to disable the automation per channel if necessary. - Evelyn Marie (leave a message · contributions) 22:41, 2 April 2023 (UTC)

All I'm asking is for you to find examples of other sources using a simple sum of subscribers as a metric. There are other ways to show the breadths of subscribers to an individuals channels than a simple sum (e.g. just list their followings on their top accounts). BrokenSegue 18:46, 3 April 2023 (UTC)

I think to put it another way, there are two extremes here. YouTuber A has two channels, both of which are subscribed to by 1 million people each - however, these 1 million people are all the same. YouTuber B also has two channels, but the 1 million subscribers to each channel are completely unique. There is no way to know which one of these, or anywhere on the spectrum between both extremes, is the case, so while its [sic] useful to have the combined total of all subscribers, without an RS to support that combined total, there is literally no way of knowing that value. Folks can add two or three numbers together if they really want to get a rough estimate of those figures. Primefac (talk) 19:07, 3 April 2023 (UTC)

I knew going in that this was a very big change to roll out, affecting thousands of infoboxes, so I expected that there would be some kinks to work out. Given those expectations, I'm actually fairly pleased at how it's gone. As anticipated, we've uncovered a few kinks in edge cases, and we're going to work those out (with patience appreciated while we do so), but it seems to be working as intended for the vast majority of instances.

For (1), the change I would support would be to have the infobox use whichever value is more recent. I wasn't able to code it to do that without my brain turning inside out, but if one of you is, feel free to go for it. I would not support having the manual entries override the automatic ones, as the entire point of this enterprise is to move away from manual entries for a piece of information that fundamentally is much better served by automatic updating than manual updating. If there are currently too many issues for the change to remain live, I'd prefer to have it just rolled back with the understanding it'll be reimplemented once we've had time to address the issues. But from the level of issues I'm seeing so far, it does not appear that there is a need to do that.

Cheers, {{u|Sdkb}} ^talk 23:44, 28 March 2023 (UTC)

I agree that seems like a sensible way to handle sub counts. An easier solution might be to just take the larger of the two values given that sub counts almost always go up? The main problem with your proposal is parsing dates consistently. BrokenSegue 02:03, 29 March 2023 (UTC)

What if I made a new lua module that implemented "newer value" logic where you'd use it like {{#invoke:NewerValue|v1=101 million|d1=March 2020|v2=111 million|d2=March 2023}} and it would return the value of v2 (in this case). BrokenSegue 02:42, 29 March 2023 (UTC)

@BrokenSegue, if that wouldn't be too hard of a module to create, then I'd say go for it! {{u|Sdkb}} ^talk 04:05, 29 March 2023 (UTC)

If we're going the bot route, then remove all option to even give a count manually. This way people aren't confused (i.e. the information will show up without parameters), and people won't make things up to get around any date values (i.e. they will very quickly figure out that putting "today's date" will keep their value ahead of the bot value). In other words, if the bot is working as intended, and there are mechanisms in place to point people where to go if there is an issue, there's no need to have a named parameter in the infobox. Primefac (talk) 10:01, 29 March 2023 (UTC)

That's where I'd like to work toward. We have to first take care of the error tracking category, though, to ensure that everything will work once the full launch is done. {{u|Sdkb}} ^talk 20:46, 29 March 2023 (UTC)

Oh, for sure, let's work out the bugs first. My suggestion was more long-term. Primefac (talk) 07:39, 30 March 2023 (UTC)

so do we want that module? I can write it this weekend if there's interest. BrokenSegue 02:46, 31 March 2023 (UTC)

@BrokenSegue The newer value one? I'd say yes! {{u|Sdkb}} ^talk 13:51, 31 March 2023 (UTC)

ok I made it. It's at Module:MostRecentValue. You can look at the test cases to see how it works: Module talk:MostRecentValue/testcases. Tell me if this is sufficient for your needs. Is there a central repo of modules that people use that I can publish this to to encourage reuse? BrokenSegue 01:14, 1 April 2023 (UTC)

Kinda not a great idea if we're having to list multiple channels. It'd be forced and against current consensus on those articles. Leave it up to the editors of the page to put in which channels to pull from. SWinxy (talk) 22:36, 29 March 2023 (UTC)

I haven't looked into how the bot/module works, but I feel like that is something that could be coded for? Primefac (talk) 07:39, 30 March 2023 (UTC)

currently the bot only ever shows a singular youtube channel subscriber count. if you want to do more than that I can add functionality for it but at the moment it's not clear how it should look. Also, the bot is only updating the sub count for one yt channel per article so that would need to change to. All of this is doable though and I'm willing to put in the time to make this happen. I just need people to say what it should look like for various cases and I can try to figure out a way to represent that in wikidata and in the module. BrokenSegue 02:46, 31 March 2023 (UTC)

Is it possible for a module to grab a specific entry in a property? I'm thinking it'd be cool to {{YouTubeSubscribers|subCountNice|UCfga98sd7AFJK7vACA|}} and it spit out the number for it. It'd be easier to make the editor make it than for the module to format it. SWinxy (talk) 17:46, 1 April 2023 (UTC)

that's possible but only if the information for that YT item is attached to the current wikidata item. So for example Wikidata doesn't think Linus Sebastian (wikidata:Q22810736) has any yt channels right now because the channels are linked to other items. But if you were willing to write something like {{YouTubeSubscribers|subCountNice|chan=UCfga98sd7AFJK7vACA|item=Q111862397}} this could be done. The wikidata mediawiki interface doesn't let you ask "find the youtube channel called $foo anywhere in wikidata and then tell me its current sub count" (that would be too inefficient). All you can ask is "for this item find me the subcount for the attached youtube channel". BrokenSegue 00:38, 2 April 2023 (UTC)

Ah yeah. I left out in my question that it would be pulling from the linked wikidata item. I think I like that a lot. SWinxy (talk) 21:59, 2 April 2023 (UTC)

Comment - Hey, I've been reading this discussion for the past few days and I wanted to ask how this module also accounts for YouTubers who have had their account terminated, along with the infobox automatically linking to a YouTube channel in question. For example, I've been working the past few weeks on the article of YouTuber Etika to get it to GA status. To make a long story short, YouTube terminated two of Etika's accounts for uploading inappropriate material between October 2018 and April 2019. However, the infobox still links to his original channel which is deleted (under the subscribers and total views section). Seeing as how there's no way to access his YouTube stats unless with SocialBlade or an archive link, I think in that case it would be redundant to have that link appear as a reference. Perhaps there should be a parameter that grants the option for users to either add or omit a YouTube channel's link from the infobox, or find a way to integrate it with an archive tool or SocialBlade functionality when a YouTuber's channel no longer exists. PantheonRadiance (talk) 23:00, 3 April 2023 (UTC)

That's a good question, @PantheonRadiance. The template currently (separately from any changes we're making here) automatically creates a reference to the channel URL for any subscribers count, and there's no way to add an archived URL. We could add a parameter for that, and edit the Wikidata entry, but I think it might be better to create a new generalized override parameter, e.g. |subscribers_manual=, that could handle any sort of edge case. That would also work for having multiple channels as Evelyn_Marie is seeking above. If that parameter is used, you could put in whatever data/reference and it'd override the bot-derived data. {{u|Sdkb}} ^talk 00:57, 4 April 2023 (UTC)

@Sdkb: This sounds promising then; adding a new parameter to manually set that up would definitely work for any edge cases, and it may be convenient for separating the stats of the channel from the rest of the other sources by grouping it as a primary source. I've noticed a bit of a pattern recently where YouTube articles have categorized sources in the References section based on whether they're primary/self-published or secondary, reliable articles, and I think having this parameter would make it easier to place stat-based info in a section for primary sources. Thank you for all your work on improving the infoboxes! PantheonRadiance (talk) 18:25, 4 April 2023 (UTC)

I've implemented 1 for a few reasons. While I don't really have an opinion on if total subscribe count is a good thing or not (it does seem somewhat WP:ORy to me), I think that a consensus should be gained on that point before overriding articles that do have that.

More importantly, there's no indication that the infobox is pulling its data from Wikidata (I think pretty much every case I've seen infoboxes pull from Wikidata at least indicates it), which makes it a very confusing situation for anyone who is not intimately familiar with Wikidata in infoboxes. The bot threshold of only updating for 10% increases also seems pretty high to me, and leads to e.g. articles like MrBeast having an infobox count older than than the article lead. The bot should at least update the count monthly regardless, so that the date in the infobox shows that the count is current as of this month (rather than a couple months old February 2023 as a lot of articles including MrBeast showed).

Also I think it's somewhat confusing having both a "Last updated" for the infobox and a separate date for the Wikidata update. Is there a reason that if we are transitioning to using Wikidata for the stats, to not have both the subscribes and the view count pull from Wikidata, and we can use the "stats updated" parameter for the date of the last Wikidata update? Galobtter (talk) 08:09, 10 April 2023 (UTC)

Category:Pages with YouTubeSubscribers module errors

Category:Pages with YouTubeSubscribers module errors is currently flooded with over 700 pages. What is causing these errors and what can we do to empty the category? wbm1058 (talk) 14:07, 5 April 2023 (UTC)

I can summarize broadly what the issues are:

1. Some items have no value for channel ID. Someone needs to lookup the channel ID for the creator's main channel and add it to the item. (you can see me do this here; unrelatedly the Wikipedia page for this person is broken and I had to look it up by going to Portuguese Wikipedia)
2. Some items have values but have no single best value. What needs to be done is someone needs to go uprank one of the several youtube channels. Then you need to wait a few days for the bot to run and fetch the most recent sub count (or you can populate it yourself manually). You can see me do this here. If there is no singular best channel then at the moment we have no fix (see above discussion).
3. Some pages don't have a wikidata item at all like Quenlin_Blackwell. An item needs to be created and the first step needs to be done.

Also, it would help if Wikidata fully supported YouTube handles. It's on my todo list to allow you to enter the user's handle instead of their channel ID on wikidata and have the bot lookup the channel ID from that (it has become quite difficult to lookup channel IDs). Unfortunately my time is a little strained at the moment. BrokenSegue 05:08, 6 April 2023 (UTC)

I have used this website to find the channel ID: commentpicker.com.  — Archer1234 (t·c) 11:04, 6 April 2023 (UTC)

This is super helpful, thank you. I'll get to work on clearing the category later today. ReneeWrites (talk) 12:31, 6 April 2023 (UTC)

OK, this is what I suspected was happening. The lack of Wikidata should never be considered a bad, top-priority error on Wikipedia. It's an omission, dammit, not an error. Wikipedia editors should not be required, or even arm-twisted, into editing Wikidata, which takes complexity to 11 for newbies who don't even understand the Wikipedia basics. Sdkb, if you or some other template editor don't promptly fix Template:Infobox YouTube personality so that it doesn't flood main namespace errors with false positives (obscuring real errors), I will. Thanks, wbm1058 (talk) 12:35, 6 April 2023 (UTC)

@Wbm1058: Ok but this isn't a "top priority" error as you said. Is there any way to mark this as an "issue" but not a major one? Nothing is actually wrong with the vast majority of these pages. I had the module mark it as an error but I can change it to issue some other warning. The point is just to have a category where all the problems live so people can go and fix them if they want to. BrokenSegue 13:00, 6 April 2023 (UTC)

I think the problem is with code like this:

{{#iferror:{{#invoke:YouTubeSubscribers|subCountNice}}|{{{subscribers|}}}|{{#invoke:YouTubeSubscribers|subCountNice}}}}

We shouldn't be testing for Module:YouTubeSubscribers to return an error, and that module should not populate Category:Pages with YouTubeSubscribers module errors when there is simply no data. It should just return a code indicating "no data". We should be testing for Module:YouTubeSubscribers to return "no data found". – wbm1058 (talk) 13:13, 6 April 2023 (UTC)

I'm not really a fluent module coder yet (still haven't managed to make time to study Lua coding more) but I see four error messages in that module:

• "Encountered a statement with zero or multiple point in time (P85) qualifiers. Please add or remove point in time information so each statement has exactly one"
• "Could not find a date for YouTube subscriber information. Is there a social media followers statement (P8687) qualified with good values for P585 and P2397?"
• "Could not find a single best YouTube channel ID for this item. Add a YouTube channel ID or set the rank of one channel ID to be preferred" (code -404)
• "Found an associated YouTube channel ID but could not find a most recent value for social media followers (i.e. P8687 qualified with P585 and P2397)" (code -412)

These messages are just disappearing into the ether. Codes should be returned to the calling module to tell it what the problem is. You might populate four sub-categories, one for each distinct problem. – wbm1058 (talk) 13:47, 6 April 2023 (UTC)

Sorry for my ignorance but why is populating four subcategories better than what is currently being done? I mean sure it's more specific but how would it be different from the current situation which you say is where it's being "considered a bad, top-priority error"? Like it'll still be adding the same number of pages to categories but just with more specific error conditions? Right? I'm not getting what about the current situation is a problem (I mean sure more specific categories would be good but that doesn't seem like the problem you are pointing out). BrokenSegue 14:15, 6 April 2023 (UTC)

What is function returnError doing?

function returnError(frame, eMessage)
	return frame:expandTemplate{ title = 'error', args = { eMessage } } .. "[[Category:Pages with YouTubeSubscribers module errors]]"
end

It appears to be populating the category, but that doesn't seem to be the only thing it's doing. It also seems to be causing the translusion of {{error}}. It should stop doing that. You might do better to just remove this function and not call it. Its only purpose seems to be to lump different issues together, making editors work to manually figure out what the issue is, when you could return specific issues without lumping them together. Whether you just return specific problem codes to the calling module or populate (multiple) categories as well, I don't care (since these are not problems I expect to work on myself). I do clear {{error}}s. When {{error}} is transcluded on a page, it should usually be accompanied by a big, bold, red error message pointing out exactly where the editor error is and giving instructions on how to fix it. For Wikidata issues, you should only populate categories, not return {{error}}s. – wbm1058 (talk) 15:20, 6 April 2023 (UTC)

ok so you don't care if we put it into a category. you just don't want me to use the error template. is there another template that shows a bolded error message but that doesn't put it into whatever group you are tracking? we currently hide the error messages but they are visible on the Module talk:YouTubeSubscribers/testcases page. BrokenSegue 02:16, 7 April 2023 (UTC)

I use {{strongbad}} for documentation of errors. See Template:Requested move#Error messages.

Someone else I raised this issue with created {{error-small}} for reporting "small errors".

I see how the {{error}} documentation led you down this path.

This meta template returns a state of error (recognized as such by a wiki parser function), and optionally an error-message text in red (visible for the user). It is used by other templates to signal an error, for example invalid input, which can then either be displayed to the editor or caught and handled by other templates.

In my view the message should not be optional, and should always be visible. Explaining my philosophy. We don't want to create hundreds of categories for unusual bad errors and make trolls gnomes trawl through them all searching for errors in haystacks. Not all error conditions populate categories. The error condition is the central location for patrolling top-priority errors. Patrolled by checking "what links here", transcludes {{error}}. OK the "error condition" also includes errors generated by template loops and parser functions such as formatting errors in math calculations, which don't transclude {{error}}, but populate categories. Common problems, like lack of sources, and lack of data, we swim in those, and organize them into swimming pools using things called {{dated maintenance category}} and {{progress box}}. The error condition should be kept to a manageable size so that emptying it can be done with paper towels rather than filtering through a pool drain. – wbm1058 (talk) 11:54, 7 April 2023 (UTC)

so I will leave the decision of whether to show the message to @Sdkb:. They designed the Template integration for the module that decides not to show the text. As far as the Module that I wrote it sounds like you would be satisfied if I switch to Template:error-small. Is that ok with you Sdkb? BrokenSegue 01:31, 8 April 2023 (UTC)

Category:Pages with YouTubeSubscribers module errors has been cleared out though I don't know how that happened. However the main namespace error transclusions are still there. I see that Template:Infobox YouTube personality is currently undergoing work by Prefall who hasn't recently been active on this talk page, but there is discussion on Template talk:Infobox YouTube personality so we have two talk pages to coordinate between. I can wait until Prefall is done with their work. – wbm1058 (talk) 23:29, 11 April 2023 (UTC)

To clarify, @Galobtter: first edited the template to prioritize the |subscribers= parameter over pulling from Wikidata automatically. They further explained their rationale on this talk page, in the section above. My changes to the template have been mostly unrelated, but I did retain and cleanup that specific change by Galobtter as I generally agree with their assessment that we should hold off on forcing the Wikidata implementation until its functionality has been expanded. Prefall 01:20, 12 April 2023 (UTC)

I'm back to debugging this. Using Jimmy Conrad as my test page, this code seems to be executed there:

{{#invoke:YouTubeSubscribers|subCountNice}}

generating the error: Module:YouTubeSubscribers:240: Could not find a single best YouTube channel ID for this item. Add a YouTube channel ID or set the rank of one channel ID to be preferred

I see this error message is on line 240 of the module.

I would like to replace that error with a return code. Something like "no-best-channel-ID-found" or whatever.

Then in Template:Infobox YouTube personality change:

{{#iferror:{{#invoke:YouTubeSubscribers|subCountNice}}}} to:

{{#ifeq:{{#invoke:YouTubeSubscribers|subCountNice}}|no-best-channel-ID-found}}

— wbm1058 (talk) 02:26, 26 April 2023 (UTC)

OK, no response after 3 weeks, so I'm assuming that everyone is OK with me proceeding to fix this myself.

I will use the following two testcases:

• Amanda Cerny as an example of a person who does have a YouTube channel ID specified in Wikidata (wikidata:Q28549924 – YouTube channel ID: UCbyGK5QKd3z0POnjEQ8jdtg)
• Jimmy Conrad as an example of a person who does NOT have a YouTube channel ID specified in Wikidata (wikidata:Q1385340)

I updated Module:YouTubeSubscribers/testcases to add these two cases.

Test results show that Amanda Cerny has 2.69 million subscribers, which is confirmed by the infobox in her Wikipedia biography.

Test results give a "Could not find a single best YouTube channel ID for this item" error response for Jimmy Conrad. It is this behavior that I'm going to change. – wbm1058 (talk) 13:03, 20 May 2023 (UTC)

I don't know why all the tests are indicated in the test results as "failing". The Cerny test seems to be successful to me, reporting "2.69 million". Why does the results page say it failed? wbm1058 (talk) 13:12, 20 May 2023 (UTC)

Because the results page is misconfigured in a way that reports all tests as failed. Personally I would abolish the separate results page and have Module:YouTubeSubscribers/testcases just be a Wikitext page containing the results in the same way as template test cases work, since that seems to be the style of testing that has been chosen anyway. * Pppery * _{it has begun...} 16:12, 21 May 2023 (UTC)

@Pppery: and now Test results all are passing! Can you tell what changed to make that happen? I don't think it's anything I've done. – wbm1058 (talk) 12:31, 25 May 2023 (UTC)

Because of a bad edit to Module:UnitTests that incorrectly marked failed tests as passed * Pppery * _{it has begun...} 13:03, 25 May 2023 (UTC)

The patch I installed today resolved the bulk of them but there are a few less-common scenarios remaining. Rather than error the code -404 is returned as marked above.

Jason Marsden (wikidata:Q739105) is returning the error "Found an associated YouTube channel ID but could not find a most recent value for social media followers (i.e. P8687 qualified with P585 and P2397)" which is another one of the four errors listed above. I added this one to the testcases. – wbm1058 (talk) 01:34, 22 May 2023 (UTC)

He does have a YouTube channel ID specified in Wikidata (YouTube channel ID: UCW8N7CjNUTsdbTzfOGuCqIQ)

Amanda Cerny has social media followers (P8687) data, and Jason Marsden has this data too

Cerny has a point in time (P585) specified for her YouTube channel ID followers (22 March 2023), while Marsden does not – wbm1058 (talk) 12:07, 22 May 2023 (UTC)

And now this one returns the code -412 as marked above. – wbm1058 (talk) 17:19, 22 May 2023 (UTC)

Learning more Lua functions as I work on this. I see that in addition to the four explicit errors I listed above, the assert function may generate three more:

• "No qid found for page. Please make a Wikidata item for this article" (code -424)
• "No such item found: "
• "Could not find a single best YouTube channel ID for this item. Add a YouTube channel ID or set the rank of one channel ID to be preferred"

That third one ("Could not find a single best YouTube channel ID...") was generated in two places. One an "error", one an "assert".

Karbin is currently generating the "No qid found for page. Please make a Wikidata item for this article" error. I just added a test with no qid specified and that duplicated this error. Again, lack of Wikidata is never a Wikipedia error. I'll patch this. – wbm1058 (talk) 13:23, 25 May 2023 (UTC)

This one returns the code -424 as marked above. – wbm1058 (talk) 23:48, 25 May 2023 (UTC)

— I've updated the Module:YouTubeSubscribers documentation to document these three new data retrieval problem codes. – wbm1058 (talk) 17:41, 26 May 2023 (UTC)

Etika - Good Article Review/Feedback

Hello everyone, I have a bit of a request. I mentioned a few days ago that I worked on YouTuber Etika's article, and a week ago I nominated it for GA status. Per this essay, I would like to ask the WikiProject if any experienced GA reviewers would like to review the article. If anyone can begin the review process within the next 7 days - from perhaps the 7th to the 14th at the latest - I would genuinely appreciate it. I will be free the entire week to work on improvements should the article be reviewed then. A couple notes for any willing reviewers:

1: After nomination, mainly copy edits have been made to the article the past few days.

2: I separated primary and secondary sources to facilitate the review process, and some of the primary/SPSs I may remove tomorrow as well.

3: Etika's history is quite a controversial topic, but virtually all disputes on the article's content have been resolved in the past few years and almost all controversial statements are backed up with fully reliable sources. At worst, it does use situationally reliable sources but in line with their descriptions at WP:RSPSS.

More notes are available on the talk page. Thanks, PantheonRadiance (talk) 08:00, 7 April 2023 (UTC)

RFC: In 2023, is Deji Olatunji notable enough to be removed from WP:DEEPER and have a Wikipedia biography?

The following discussion is an archived record of a request for comment. Please do not modify it. No further edits should be made to this discussion. A summary of the conclusions reached follows.

Normally this wouldn't merit a formal closure, but it was requested and with the history it's not an unreasonable ask. There is consensus to remove them from WP:DEEPER and to allow work on an article. I'm not going to say there's a consensus to have a Wikipedia biography, because this isn't where that is decided. ScottishFinnishRadish (talk) 15:07, 22 June 2023 (UTC)

---

In 2023, is Deji Olatunji notable enough to be removed from WP:DEEPER and have a Wikipedia biography? Is Deji notable enough for an article in 2023? DrewieStewie (talk) 21:43, 31 May 2023 (UTC)

As you can see, this proposal aims to discuss the notability of Deji Olatunji, KSI's brother formerly known as ComedyShortsGamer, who has 10 million subscribers on YouTube and has expanded to several other ventures. As a little bit of background: several times in the mid-to-late 2010s up until March 2022, there were more than 20 deletions of articles for Deji under G4 and G5, along with at least two AFDs then, including this one. Around October and November 2022, another article (under the title Deji (YouTuber)) was created, due to his fame having significantly grown in recent years following the speedy deletions and AFDs, to the point of facing Floyd Mayweather Jr. in an pay-per-view exhibition boxing match. However, in the midst of a move request discussion initiated by me, it was deleted under G5, despite there (at the time) being no voiced objections to the subject's notability or any initiated AFDs. Afterwards, I established the most recent deletion review, which had two camps forming both in support and opposition of his notability. Ultimately, the closer decided that there was a consensus against his notability, citing User:JzG/And the band played on.... Subsequently, he was added to WP:DEEPER. Afterwards, there was dissent from me and User:PantheonRadiance on User: Sandstein's (the closer) talk page (see talk page discussion here), with the proposal to establish this RFC to discuss it further. However, I didn't start it until now, due to being preoccupied with my education and to give the issue some rest temporarily.

Opponents to his notability argue that good faith has been exhausted for the subject, due to the 20+ G4 and G5 deletions. They note that User:JzG/And the band played on... applies, even though the subject, Deji, never proposed the article himself or encouraged the behavior of the Ultras (in this case referring to his hardcore fans) that continued creating poor-quality startup articles for the subject. They believe that this erosion of good faith overrides any increased notability after the fact.

Dissenting voices who believe Deji is notable enough for an article believe that his notability exceeds by a long shot that of the other entries at WP:DEEPER. They believe that his reliable source coverage is significant, and that any Wikipedia stigmas caused by the ultra editors are very unfair to the subject. They believe the latest drafts at the time of the Deletion Review, as well as the latest article at Deji {YouTuber) which was deleted just before the Mayweather vs Deji bout, were satisfactory articles that established notability for the subject. They believe that notability can change and elevate to Wikipedia levels later, even in the face of 20+ G4 and G5s beforehand. PantheonRadiance stated it best on Sandstein's talk page:

The deletion did have quite a lot of merit in the past as many of the previous versions were riddled with unreliable, primary, and/or non-independent sources, if they weren't stacked to the brim with original research. Based on those, community consensus was perfectly justified in deleting the article. However, looking at the recent versions that were created by blocked users, it was clear that there were plenty of reliable and significant coverage of Deji between his video content, his boxing career and personal life. I was going to post my two cents on the article at the deletion review too but it closed before I got a chance to do so. But basically put, there were reliable sources that significantly covered him even present in some of the AfDs (this one for example), that were dismissed because they were "trivial" even though they did explain significant aspects of his YouTube career and life that could've flashed out the article. Not only were sources like BBC and The Daily Dot present that covered Deji, among others from Business Insider and The Independent, but recently Deji's boxing match received coverage from Sky Sports, ESPN and Bleacher Report. Even considering the poorly made attempts at creating this article in the past, I think even the most cynical Wikipedia editor can't seriously dismiss all of these sources as trivial coverage. From a quantitative standpoint each source covers him in multiple paragraphs as opposed to passing mentions. And at best, saying the topic of these articles is insignificant is merely subjective skepticism that doesn't change the fact that reliable media outlets consider him significant enough to report on him. And at the end of the day, isn't that all that should matter?

Therefore, I once again ask in this RFC: Is Deji notable enough for an article in 2023? DrewieStewie (talk) 21:43, 31 May 2023 (UTC)

• Yes, as poster, per my various comments on the matter and per User:PantheonRadiance's comments on the matter/ability to establish Wikipedia notability for YouTubers as well. DrewieStewie (talk) 21:45, 31 May 2023 (UTC)
• I'm extremely skeptical of this idea, but I suppose no harm can come from permitting a draftspace only page to be created and go through WP:AFC process so we can avoid the umpteenth ANI fight.  — SMcCandlish ☏ ¢ 😼  23:58, 31 May 2023 (UTC)

  I'd second the draftspace page. I don't think an RFC is the appropriate mechanism for this, and, frankly, with respect, this RFC doesn't comply with WP:RFCNEUTRAL or WP:RFCBRIEF.--Jerome Frank Disciple 22:50, 3 June 2023 (UTC)

  Deji had already been through multiple drafts that were declined and rejected in the past, many of which from ultras as pointed out by DrewieStewie. One of the final drafts that existed during the DRV in 2022 was deleted per WP:G6, shortly after the DRV consensus emerged that he doesn't merit an article. I'd definitely say that an RfC is warranted in order to prove once and for all that he is unambiguously notable, lest this becomes so controversial that any editor who tries to create a page or even a draft receives any sort of punishment under assumption of bad faith. I'm still in the midst of finding sources that prove his notability, but for now, I'll leave some that do cover him significantly:

  The Verge: 1

  The Independent: 1, 2

  BBC News: 1

  Kirkus book review (potentially meets WP:NAUTHOR?): 1

  PantheonRadiance (talk) 02:45, 4 June 2023 (UTC)

  The only problem with the RFC is that the nominator failed to sign immediately after the question so the bot thinks his (allowed by policy) arguments for his position are part of the question... which they clearly aren't. The arguments are clearly separated from the neutrally worded question by a paragraph break. Calling out WP:RFCNEUTRAL and WP:RFCBRIEF is a bit much for what amounts to a mistake akin to a typo. Fieari (talk) 04:29, 5 June 2023 (UTC)

  Appreciate the correction; that was my mistake that I had failed to notice earlier. DrewieStewie (talk) 05:16, 5 June 2023 (UTC)

• Yes, I haven't been involved in this discussion in the past, reviewing this person there seems to be plenty of coverage to justify a biography article. Nemov (talk) 02:57, 1 June 2023 (UTC)
• Not sure we need an RfC - why not complete a suitable draft to show it meets GNG and go through AfC? Lee Vilenski ^{(talk • contribs)} 08:14, 1 June 2023 (UTC)

  @Lee Vilenski: Hello my friend. Good to see you here. RFC is needed due to several reasons. First, there’s the salting of several article titles relating to subject (hence the WP:DEEPER inclusion). Second, there’s the risk of it being shot down out of pocket due to community bad faith existing towards the subject due to past poor behavior from fanboys attempting to create the article, hence why consensus beforehand is wise. DrewieStewie (talk) 10:56, 1 June 2023 (UTC)

• Yes - I've done a source review on the sources provided above to see if they meet WP:GNG. Remember, the criteria for WP:GNG is that there exists independent (the subject neither wrote, nor paid to have someone else write) secondary (with analysis and synthesis, not just being the source material) reporting on the subject that is significant (not a passing mention) and that covers more than just a one time single event (so coverage of lots of things is more notable than if multiple sources are all reporting on the same incident). I'd also say that more weight should be given if the coverage itself explains that the subject is notable. Here's what I found:

Reference Number	Reference	Independent	Significant	Reliable	Secondary	Notes
1	The Verge: 1	Yes	Yes, article is extensively about the subject	Yes	Yes	Topic of article is about boxing event.
2	The Independent: 1	Yes	Yes, article is about the subject (in a negative light) and even flat out identifies the subject as "among YouTube's biggest stars"	Yes	Yes	Topic is a feud.
3	The Independent: 2	Yes	Yes, article is extensively about the subject	Yes	Yes	Topic is a feud.
4	BBC News: 1	Yes	Yes, article is about the subject	Yes	Yes	Topic is about a legal matter involving subject's dog, which would not have been notable to write about had not the subject been notable.
5	Kirkus : 1	Yes	Yes, review is not only about the book the subject wrote but also covers a bit about the subject himself	Yes	Yes	Main topic is a professional review of book the subject wrote. (the reviewer is not kind)

These sources definitely look like they meet WP:GNG to me. I'm not sure how you could claim they don't. I don't know this YouTuber, I'm not a fan, and I have no dog in this fight. I'm just looking at our policies and the coverage provided, and I see no reason for this person's article to be deleted, whether by the AFD process or by speedy deletion. Fieari (talk) 04:55, 5 June 2023 (UTC)

• Yes - Based on the evidence provided in recent comments, it seems the subject has become notable enough to merit an article. JoseJan89 (talk) 07:13, 7 June 2023 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Requested move at Talk:Daithi De Nogla#Requested move 13 June 2023

There is a requested move discussion at Talk:Daithi De Nogla#Requested move 13 June 2023 that may be of interest to members of this WikiProject. Captain Jack Sparrow (talk) 16:47, 20 June 2023 (UTC)

Help with declined draft

Draft:Krew (Youtube group)

The reviewer said the submission is not adequately supported by reliable sources. Any help finding things that should be sourced but aren't, or unreliable sources if there are any, would be greatly appreciated. I did make some changes since the draft was declined, so if it is now good enough tell me that too. AKFkrewfamKF1 (talk) 18:54, 22 September 2023 (UTC)

Likely notable group. TVovermind is a listicle and doesn't seem reliable. THR is good, Globe and Mail I think is good, Insider is good. Maybe cut out the 'Competitions' section. Look at articles in Category:YouTube groups, e.g. Smosh, for referencing style and tone. SWinxy (talk) 19:34, 22 September 2023 (UTC)

thanks. I removed the listicle reference and competitions section. Also added more references overall, especially where there was unsorced content. Will look at articles before submitting. I did look at the Dream (YouTuber) article some to know how to do things, but I will look specifically at refferences of group channels this time. AKFkrewfamKF1 (talk) 20:13, 22 September 2023 (UTC)