Talk:Analytics/Metric definitions

"Countable namespaces"

The blog post links a non-existing section about them, anyway I suggest to drop this very confusing term and use Content namespaces instead. See bugzilla:35198. Thanks, Nemo 21:45, 31 August 2012 (UTC)

The anchor link in the blog post has been fixed, thanks for pointing that out. Regards, Tbayer (WMF) (talk)

Now there's this confusing sentence in the introduction:

In the context of wikistats edits are updates to countable pages in countable namespaces (aka mainspace).

but "Content namespaces" goes on to describe other namespaces besides the main namespace. I changed the sentence but I have no idea what's intended. -- S Page (WMF) (talk) 18:49, 13 May 2013 (UTC)

Technically speaking, "countable namespace" doesn't mean anything; it's been recently invented by the analytics team. Countable pages (for the "article count") are pages that meet the count method and are in content namespaces; [countable] edits are just all edits in content namespaces. However, the WMF has not yet aligned its stats to content namespaces: it's just a bug, but carelessly "fixing" it would make some stats hard to compare, so until the things are unified they're using "countable namespaces" to define the weird ad hoc mix used by WMF stats. :) Or at least, this is my understanding. IMHO it would be easier to just drop this useless distinction, don't bother about "rewriting history" and update all counts, translating all strategic goals to relative rather than absolute amounts so that count errors of 2010 become irrelevant. --Nemo 22:18, 13 May 2013 (UTC)

Sorry, but I still do not understand what the word "article" or "edit" referes to, that is which namespace? Does statistics only count the main namespace or does it count the templates, discussions etc. as well? (please excuse my bad english, its not my mother tongue)--Loiosch (talk) 08:34, 21 July 2015 (UTC)

Confusing definitions for page views

From the definitions:

• The url in a logline contains /wiki/. This excludes /w/index.php? and SpecialPages.

What about special pages that contain /wiki/ in the url?

• Any article namespace qualifies (unlike the dump based reports (see 'Content' above)) .

In MediaWiki, an "article" namespace is sometimes used to mean any namespace with an even non-negative namespace id. However, that definition is not used very often. Is that what is meant here? Bawolff (talk) 03:41, 7 January 2014 (UTC)

Special pages are indeed included in the raw data, I didn't check if there's some additional filtering on WikiStats (there should be for at least some like [1]).

The second should just mean content namespaces, but I don't know if the API-based definition is already in use by WikiStats for this piece too. --Nemo 06:40, 7 January 2014 (UTC)

Clarificaton needed

I have several questions about how Wikistats counts articles:

1. The "Countable pages" section says, "most Wikistats reports are based on so called stub dump[s]" — does this include the article counts shown at, for example, stats:EN/TablesArticlesTotal.htm?
2. If so, what determines whether a "full" or "stub" dump will be used for each wiki? Is there a list somewhere?
3. When "stub" dumps are used, is the effective defintion of an article simply "all pages in all content namespaces"?
4. Are the "historical" article counts recalculated every time, whether or not the dump used is "full" or "stub"?

I only really need #2–4 answered if the answer to #1 is "yes". - dcljr (talk) 22:18, 19 April 2015 (UTC)

Statistics for past months are always recalculated entirely, except for the sections of wikistats which have an archive of course. --Nemo 22:58, 19 April 2015 (UTC)

Wikipedians - What does it mean?

According to the definitions "Wikipedians" is the nickname for user with more than 10 edits. But in the table users with more than 5 edits are counted. Besides, does the number ">5" include the number >100" edits? In May 2015 for example would it be correct to say Wikipedia had 2034672 registered editors among whom 73123 made more than 5 edits among whom 10064 made more than 100 edits. (please excuse my english, its not my native language)--Loiosch (talk) 08:34, 21 July 2015 (UTC)

Found answer in https://stats.wikimedia.org/EN/TablesWikipediaZZ.htm#editor_activity_levels but it would be nice, to write that in the definitions as well. --Loiosch (talk) 08:41, 21 July 2015 (UTC)

Everything is written already: this page is not concerned with levels other than 5/month and 100/month; "5 or more edits" is very clear, X >= 100 >= 5 for the transitive property.

The status of "Contributor" has nothing to do with the status of "Active editor". Contributors is in the first table, not in "Edit activity levels of registered users". In May 2014, Wikipedia had 72841 "active editors", of which 9689 "very active"; it also reached a grand total of 1861459 "contributors", who may have ever been "active editors" or not. --Nemo 09:15, 21 July 2015 (UTC)

Move to Meta

Just a note that this page (among other things) was discussed in an October 2015 thread on the analytics mailing list, and there was general consensus to consolidate this information into the Research namespace on Meta. I plan to implement that at some point.—Neil P. Quinn-WMF (talk) 21:48, 18 January 2016 (UTC)

Wikistats 2 Metric Definitions

Wikistats 2 has the unenviable job of presenting metrics from both Wikistats and the Metric Standardization project. The new UI links to metric definitions and provides a blurb. The question is, how to best do this, where to link, etc. My first thought, and what I'm about to do right now, is to link Standard metrics to their existing Research:Metric_name pages, and create new Research:Wikistats_Metric_name pages for the new metrics. I will incorporate whatever is relevant from this page, however, some of the metrics have been expanded to be a bit more dynamic. For example, Active Editors, Very Active Editors, etc. have been merged into just "Editors" with different activity levels.

Here are the pages I'm stubbing out for now, let me know what you think and if you have different opinions (the UI will be updated with whatever links we decide here, but for now a bunch of the links are broken):

• m:Research:Wikistats_metrics
• m:Research:Wikistats_metrics/Bytes
• m:Research:Wikistats_metrics/Edited_pages
• m:Research:Wikistats_metrics/Editors
• m:Research:Wikistats_metrics/Edits
• m:Research:Wikistats_metrics/New_pages

Please feel free to disagree, comment, etc. — Preceding unsigned comment added by [[User:{{{1}}}|{{{1}}}]] ([[User talk:{{{1}}}|talk]] • [[Special:Contributions/{{{1}}}|contribs]])

Moving to Meta makes sense, but it should still be easy to find the traditional/most important metrics, while not duplicating existing pages. For instance, there should be something that m:Active editors will redirect to, and something coordinating well with Manual:Article count. Also, the content of this page may be moved to m:Research:Wikistats_metrics entirely, to keep the definitions which are relevant to WikiStats 1 and to all the subpages, such as the definition of content namespace and the treatment for bots.

It's a bit weird to have 3 pages for "bytes", "edited pages" and "new pages" and just one for "editors", but that's not so important if people can find what they're looking for. (It's often easier to start from a single page and then split the sections to their own pages when you're sure about the structure.) For "editors" there are also the lists of active and inactive registered editors, list of prolific unregistered editors and bots, distribution of editor activity, editor activity levels by namespace: that page may become quite big.

It's not clear where you'd put information on article count, not to mention number of words and so on if they were implemented. Maybe you'd benefit from more inclusive titles, such as "Page counts" (for "edited pages", "new pages", article count and whatever came up), "Total editors" (for the number of active, very active, unregistered editors etc.), "Contribution size" or whatever. Or just have a subpage for every small thing, I guess. --Nemo 08:13, 4 January 2018 (UTC)