[04:23:36] halfak, want a funny story? [04:23:48] so you know the ISO 3166 Alpha-2 standard? Standardised codes for each nation-state? [04:23:50] Sure. [04:23:52] And you know Namibia? [04:24:00] And you know how Namibia's Alpha-2 code is "NA"? ;) [04:24:14] Oh god [04:24:21] yuuup [04:24:23] lol [04:24:29] and you know how MaxMind thinks quoting their fields is for suckers? [04:25:10] na.strings = "" [04:25:16] for read.table [04:25:57] yeah, I went for "I only have to read this thing once, we're fine" [04:25:58] in the end [04:54:14] whee [04:54:24] who got the country visualisation working with dynamic maps?! [04:54:33] I DID (LL helped a helluva lot with the thunking, as did leila ) [15:44:20] o/ Ironholds [17:40:56] morning [17:41:22] hey halfak :) [17:41:27] Hey dude. [17:41:54] how goes? [17:42:09] good morning aaron and holder [17:42:29] hey hare j :) [17:42:39] harej, halfak , http://datavis.wmflabs.org/where/ - it has maps nao \o/ [17:42:58] maps maps maps maps [17:43:11] That's a weird projection you're using. It looks all... globey. [17:43:21] the mollweide? [17:43:44] yeah, it emphasises shape accuracy [17:43:47] I see. [17:44:22] that is, accuracy in shape size [17:44:28] to avoid the "europe is the size of Africa" problem [17:44:46] I was going to drive to West Virginia today, but apparently they're going to get pounded with snow starting this afternoon going into tomorrow morning. So now THAT won't be happening. I was looking forward to driving a car, too! [17:45:40] Ironholds, cool! I just showed off your stuff to the other hackers. [17:45:46] o/ harej [17:46:09] halfak, what did they say? [17:46:22] and it's not all my stuff, LL gets half credit [17:46:40] Cool. Also, we expressed sadness that we couldn't see cool things in africa due to the anonymization thrshold. [17:46:56] Wait.. who is LL? [17:47:13] halfak, actually, you'd still not see cool things in Africa [17:47:23] you should see my editing projections :( [17:47:35] and, LL - http://sarahlaplante.com/ [17:48:02] She's a Python/Java big data developer who likes search and multilanguage NLP [17:48:04] we should hire her [18:03:22] halfak, also, do you have any opinions on the user agent thread? More voices would be good (at the moment it's just me and Nuria which doesn't scream 'this is an important release' although RStu wants it [18:20:54] tnegrin, new pageviews definition is done and fully implemented. [18:23:32] Ironholds: congrats. do I understand correctly that, in the not-too-distant future, stats.wikimedia.org will be a source for page-level traffic data? [18:24:06] ragesoss, well, analytics engineering still needs to implement that and wrangle apps into doing their part, because their infrastructure won't be countered otherwise [18:24:26] in the short-term, it means we will have high-level numbers that are not painfully, PAINFULLY inaccurate. [18:24:49] in the longer-term, yes, hopefully highl-evel numbers that break out the access method and also better per-article data [18:25:12] and an API for getting per-article data, I hope? [18:25:53] no idea! I'm not working on that project any more [18:26:21] you want to ask Toby or Kevin [18:26:56] * ragesoss nods [18:27:06] page-level stats? finally???? [18:27:12] no more swedish guy? [18:32:24] harej, if AnEn gets the time and resourcing they need [18:32:38] if you want it I encourage you to poke the people above me on the analytics mailing list [18:44:52] people keep emailing me calling me Dr Keyes [18:45:07] sorry halfak, already got a PhD in the minds of the public, don't need another one ;p [18:45:18] and it only took me three years! And /they/ paid /me/. Best kind of PhD [19:01:41] hee hee [19:01:54] Jenny B and I are adding compliments to the unittest library [19:01:57] * Ironholds cracks knuckles [19:02:07] time to see if I can get the entire chorus of Katy Perry's "Firework" +2d. [19:16:59] Ironholds, trying to ignore email. Is it time sensitive? [19:26:45] halfak, huh? oh, no! Sending /me/ emails ;p [19:36:23] harej, you missed some fun conspiracy theorising [19:36:47] in another irc channel i was talking about starting a mediawiki government contracting business [19:44:18] harej, ahh, the government-wikipedia industrial complex [20:04:07] Ironholds, o/ [20:04:44] halfak, coming! [20:04:47] kk [21:38:03] Ironholds: hi! [21:38:04] morning [21:38:16] hey YuviPanda [21:38:43] Ironholds: I saw some murmurs about making some data available to labs on labs-l... [21:38:45] err [21:38:48] on analytics-l [21:39:00] yup [21:39:04] by way of sticking it at a public location [21:39:40] let me try to find the exact thread [21:39:53] (I only have digest mode) [21:40:21] well, I know the thread ;p [21:41:52] Ironholds: yeah, am reading through because I only glanced at it. [21:42:09] Ironholds: so you have a script that generates this data, ja? [21:42:49] it's nowhere near production ready, dude [21:42:54] like, this is a premature conversation [21:42:59] ah [21:42:59] I see [21:43:01] alright then [21:43:03] * YuviPanda ignores [21:43:07] when analytics engineering have built any of the systems people need, then we'll have something reliably [21:43:20] at the moment we have an ad-hoc script I wrote. This isn't enough; it needs to be an AnEng driven effort. [21:43:23] I do like getting more data out onto people’s hands [21:43:46] and going from ‘script’ to ‘data available on labs NFS’ is something I can drive from start to end [21:44:12] totally [21:44:15] but we don't have the script [21:44:26] we have a pile of R. We need a pile of Java on an Oozie job. [21:44:31] right, right [21:44:42] well, I guess you are just hitting checkuser and sampled logs. [21:44:48] which is better than nothing.. [21:55:49] yes, but not robust [21:55:59] nor maintainable [21:56:06] if, as a wild example, I stop working for R&D [22:02:25] Ironholds, just read through the ua thread. I don't see what there is to comment on. [22:02:54] halfak, just: can you see any privacy risk? would you like it if this data was released? [22:03:07] I see. Let me be critical of that bit. [22:05:41] halfak, cool! Thanks :) [22:05:55] I'm distinctly reviewing with legal at dartar's suggestion, so hopefully they'll catch any stupid as well. [22:05:59] crowdsourced stupid-catching! [22:16:18] halfak, thanks for the notes!