NIWA Stats

[Vellidragon]

What would "editors" include? All registered users or only those with some sort of activity? Just wondering, as there's an "Active users" page with "users who had some kind of activity within the last 91 days" among the special pages. Counting these manually, WiKirby currently has 55, if anyone cares.

(I'm probably completely missing something here, so I apologise for my childish ignorance.)

[Greenpickle]

Active users sounds like a good measure of the number of editors.  You can easily grab it from the MediaWiki API.  For example, you can get the general statistics, including active users, for Pikipedia, at:

http://www.pikminwiki.com/api.php?format=json&action=query&meta=siteinfo&siprop=statistics

You can fetch and parse this for active users fairly easily - for example, in JavaScript for non-IE:

Code Select Expand


var api = 'http://www.pikminwiki.com/api.php';
var r = new XMLHttpRequest();
r.open('GET', api + '?format=json&action=query&meta=siteinfo&siprop=statistics', false);
r.send(null);
alert(JSON.parse(r.responseText).query.statistics.activeusers);


Or in Python 3:

Code Select Expand


from urllib.request import urlopen
from json import loads
api = 'http://www.pikminwiki.com/api.php'
print(loads(urlopen(api + '?format=json&action=query&meta=siteinfo&siprop=statistics').read().decode())['query']['statistics']['activeusers'])


[Xizor]

I was thinking just registered members, but active users is probably much better!

Edit: Btw, Zelda Wiki caps it off at 30 days, so there needs to be some consistency between Wikis if we're to use this a valid, uniform stat, but anyway, within the last 30 days, 121 people have edited ZW.

Also, another fun stat to compile could be the total article count of all of NIWA.

[Greenpickle]

Pretty sure it's 30 days at every wiki.

[Vellidragon]

Pretty sure it's 30 days at every wiki.

"users who had some kind of activity within the last 91 days"

That's an exact quote from the Active Users page on WiKirby, so evidently, it's not 30 days everywhere.

EDIT: Just checked random wikis, and Wars Wiki, for instance, also has 91 days. There's definitely wikis with 30 as well, though.

EDIT2: Lylat Wiki and Starfy Wiki also have 91 days. The rest all seem to have 30, unless I forgot one.

[Greenpickle]

Hmm, okay.  Interesting...

[FlyingRagnar]

Looks like it should be controlled by $wgActiveUserDays.  30 is the default.  Not sure why it would be set to 91 for any wiki, weird.

[porplemontage]

Looks like it should be controlled by $wgActiveUserDays.  30 is the default.  Not sure why it would be set to 91 for any wiki, weird.

Older versions of MediaWiki.

[FlyingRagnar]

Yeah that's what I figured, but looking at the history of the article, it never lists a value different than 30.  Probably undocumented in earlier versions or something. 

[FlyingRagnar]

http://pastebin.com/Ww6MmvPR

Here is a python script I threw together based on Greenpickle's suggestion above.  It grabs various statistics from each wiki and puts it into a very basic html table.  A couple things...

1) Can one of you with better web hosting put this on a cron?  My hosting only has python 2.4 which does not support the JSON library (2.6+). 

2) I could not find the api.php for the Wars Wiki.  Seems like it should be http://www.warswiki.warscentral.com/api.php.  So that wiki is currently missing. 

The script does not do any sort of error handling at all.  The value for active users it spit out for me (minus Wars wiki) was 1893!

[Greenpickle]

For (1), since you're fetching from trusted sources, and you're quite sure the output will be valid JSON, you can just eval() it.

Quite important: you're fetching and parsing the page for every category, but you only need to do it once per wiki.  This obviously makes it run about 6 times as fast.

As for errors, if you weren't sure, urlopen gives IOError when the page is dead (or maybe it does so on .read()), and the API will give errors in the returned data structure (in the 'error' key, I think?).  Checking a page is valid JSON kind of conflicts with what I said for (1), so maybe you could have a hack like checking for fetched_string.startswith('{"query":') ?

(...And I'm struggling to refrain from complaining about your HTML - I'm telling myself it doesn't matter.)

Edit: looking at the difference between this table each month might be a better way of getting the monthly edits, too.

[FlyingRagnar]

Haha, you are so critical.    This was just 10 minutes of work, and I never write Python.  Feel free to make some quick changes and post an updated script.  I didn't work on it longer because I figured out that my hosting doesn't have Python 2.6+, so I kinda quit then. 

I will work on it again when I have some time. 

[Greenpickle]

Hey, the only point of criticism was the bit about only fetching once per wiki, and that's pretty important; the rest were just general thoughts.

[FlyingRagnar]

Oh yeah, I totally understand.  I'm thinking the cron would only run once a day, more than that and the stats probably aren't more meaningful.  So even if it was a slow process, it wouldn't matter much.  Of course, things are still thrown off by the fact that some wikis have 91 days for active users and others 30.  There are probably other config settings that skew the values.

[FlyingRagnar]

NIWA Statistics
Wiki Name	Articles	Edits	Media	Users	Active Users	Admins
Pikipedia	597	37422	1737	318	8	9
Bulbapedia	21159	1647133	0	62442	832	9
Donkey Kong Wiki	900	11664	2334	465	14	5
Fire Emblem Wiki	872	23738	6402	237	15	7
F-Zero Wiki	218	9435	332	58	16	8
Golden Sun Universe	1421	50293	4014	855	12	5
Icaruspedia	271	4968	620	112	56	6
Lylat Wiki	164	38914	1050	269	16	11
Metroid Wiki	1521	41021	2682	1813	17	10
Nintendo Wiki	518	14010	611	302	8	9
Nookipedia	1355	28669	3961	1237	34	10
Pikmin Fanon Wiki	1647	26338	1804	231	14	7
Smash Wiki	2068	105067	5983	1194	88	9
Starfy Wiki	725	13535	1577	92	12	7
Strategy Wiki	23559	598018	80586	60008	78	32
Mario Wiki	12121	1272908	36407	23348	489	12
Wikibound	977	20181	2415	569	13	4
Wikirby	740	19385	2175	668	45	4
Zelda Wiki	3629	301538	13834	9763	121	17
Dragon Quest Wiki	2657	29724	4937	287	15	3
Totals	77119	4293961	173461	164268	1903	184
2012-04-03 12:40:03.384798

Not sure why Bulbapedia has no media/images.  Maybe because they are using a customized storage system for their files?

[Moydow]

Bulbagarden and Bulbapedia use Bulbagarden Archives for media storage for the entire site instead, similar to Wikimedia and the Wikimedia Commons.

[Bop1996]

Metroid Wiki had a fairly good month compared to what activity has been like lately. Of those 17 active users, three were admins, and most of the rest were people that edited once or twice and then didn't edit anymore, tho.

[Greenpickle]

Are you still going to be gathering data for the number of edits per week?

[FlyingRagnar]

Wiki Name	Average Posts Per Week	Percentage	Change
Bulbapedia	4342.3	36.83%	-1.95%
Donkey Kong Wiki	111.5	0.95%	-23.68%
Dragon Quest Wiki	38.5	0.33%
Fire Emblem Wiki	371.9	3.15%	+54.57%
F-Zero Wiki	94	0.80%
Golden Sun Universe	76.8	0.65%	+3.50%
Icaruspedia	226.6	1.92%
Lylat Wiki	8.4	0.07%	+140.00%
Mario Wiki	2806.3	23.80%	+8.13%
Metroid Wiki	190.9	1.62%	-3.05%
Nintendo Wiki	47.1	0.40%	+20.77%
Nookipedia	60.9	0.52%	-13.62%
Pikipedia	69.3	0.59%	+10.35%
Pikmin Fanon	49	0.42%	-2.39%
SmashWiki	1165	9.88%	+2.48%
Starfy Wiki	474.4	4.02%	+29.02%
StrategyWiki	765.3	6.49%	+43.05%
Wars Wiki	71.6	0.61%	+38.76%
Wikibound	102.2	0.87%	+35.54%
WiKirby	40.1	0.34%	-35.11%
Zelda Wiki	678.1	5.75%	-32.54%
Total	11790.2	100.00%	+5.82%
Avg Posts Per Week	561.4380952		-9.29%
Date	4/3/2012

[Toomai]

Here, have some graphs: