[12:22:08] halfak: Guest12037 do you know if Maximilian Klein is on IRC? [12:22:16] oh [12:22:35] it’s notconfusing [12:22:36] ok [15:21:10] morning halfak [15:21:16] Hey Ironholds [15:21:20] how goes? [18:41:10] Nettrom, so. [18:41:14] yes [18:41:27] I was thinking about articles that have short term popularity vs. long term popularity. [18:41:41] And how inlinks seems to be a good predictor of view rate... [18:42:36] And I thought that maybe you could detect sustained vs. short-term view rates without waiting to see what happens by using other predictors like that. [18:43:05] So, if an article has a high view rate, but a low inlink count, I hypothesize that the high view rate will be short-lived. [18:43:14] mhm [18:43:17] I'm not sure how this would work out with breaking news events though. [18:43:31] I suspect they would tend to get linked a lot. [18:43:38] But then again -- maybe not. [18:44:07] sounds like this needs a bit of neighbourhood analysis, similar to the geomodels we're doing [18:44:24] neighborhood? [18:44:42] I had a quick look at articles related to Robin Williams, and when he did, of course his article got a lot of hits, but so did e.g. articles about his movies [18:45:07] so I'm thinking that one would have to do spatial autocorrelation based on the link network [18:45:28] "spacial" seems like the wrong word. [18:45:30] hmmm... wouldn't that simply correlate with inlink count, though? [18:45:34] But I get what you are saying. [18:45:59] *spatial [18:46:02] * halfak feels shame [18:46:14] pft, space shmaze [18:46:16] :D [18:46:43] hmmm, I'll keep this in the back of my head [18:46:48] So, regardless of nuance that you could bring to the analysis, I bet simple predictors would do just fine. [18:47:05] Cool. I'll have link data for you in an easy to consume form when you are ready for it. [18:50:27] thanks! not sure if it'll make it this time around, short deadline, but good to think about if I'm looking for more robust measures down the line [19:58:20] is anyone else having trouble SSHing into stat1002? [20:43:28] Ironholds, just saw your question. Stil having trouble? [20:43:53] halfak, seems to have resolved :) [20:44:04] kk [21:03:59] * Ironholds giggles [21:04:12] halfak, most of the WikipediR documentation is now written around your article. YW. [21:04:25] my article? [21:04:41] #'@examples [21:04:41] #'#Backlink [21:04:42] #'all_bls <- page_backlinks("en","wikipedia", page = "Aaron Halfaker") [21:04:48] lol [21:05:00] (actually did laugh out loud) [21:05:16] Are there any backlinks? [21:05:41] yup! [21:06:11] "John T. Riedl", "Criticism of Wikipedia", "GroupLens Research" [21:06:29] Oh god. My age is in there now. [21:10:39] halfak, well, if I hadn't found out when your birthday was.. [21:11:17] How'd you get the year? [21:11:26] you told me you were 31, and I knew the day, so I counted [21:11:43] a year as a research scientist has taught me one thing, and one thing only; basic arithmetic [21:12:17] heh. {{citation needed}} [21:12:21] :P [21:12:30] oi! [21:12:38] * Ironholds goes to add mean things to halfak's article [21:13:06] "Once convicted of cruelty to animals due to a scenario involving seven ferrets, a 50-gallon drum of marmalade and a frat house" coming RIGHT up. [21:14:22] It's important to have seven ferrets so that you can lose a couple and still be able to complete the pentagram,. [21:14:32] exactly [21:14:44] and also in case any of them get sick and tired of marmalade-flavoured bro [21:14:48] you don't want some bros to get out un-nibbled. [21:14:50] that's just a waste. [21:16:20] * halfak is patiently awaiting the deletion discussion [21:16:29] Or maybe I'll survive and have a revert war. [21:16:34] hey Ironholds -- did the pentaho data get uploaded? [21:16:54] it's still being generated [21:17:10] ok cool -- just looked at the cube and saw it was off -- np [21:17:11] one of the disadvantages of fixing a bug that consists of accidentally filtering too much is now there's more data to parse ;p [21:17:30] halfak, I mean, *I* think you're important. [21:17:43] and people tend not to touch my articles [21:18:45] I saw your COI template on the talk page. I feel like there should be a special one for the subject of the article. One that sets off alarms if I even add it to my watchlist. [21:20:53] haha [21:22:05] I always get confused for a second when I remember that halfak's last name isn't actuallyh "Halfak" [21:22:55] heh. My last name is just descriptive of my IRC handle. Halfaking would be what I'm doing when I use the name "halfak", therefore, I am a halfaker. [21:23:10] A. Halfaker, that is. ;) [21:24:28] :> [21:24:31] Yeah, I mentally map you to "Aaron" or "Halfak" [21:24:41] but people call me "Ironholds" in meatspace [21:27:50] Ironholds, I regularly have an urge to do so. [21:27:55] haha [21:28:06] and J-Mo is just J-Mo, but in my defence I never called him Jonathan [21:28:08] I sometimes have to stop and think for a tick [21:28:17] he migrated to J-Mo from Little Johnny Morgan, because it's an adorable name for an adorable person [21:28:20] +1 J-Mo is an excellent nick [21:28:29] and I think J-Mo is profoundly grateful that it resolved to J-Mo in the end [21:28:53] also: staff! [21:29:09] yep. that's my name. don't wear it out. [21:29:26] When people ask me, "Have you read the work about the teahouse? I think the author is John Morgan or something." "Oh! Yeah! I know J-Mo. Huge fan." [21:29:48] that happens often, does it halfak? [21:29:55] In fact, it does :) [21:30:07] * J-Mo pats himself on the back [21:31:38] STAFFY STAFFER McSTAFFERTON [21:32:53] DarTar, Ironholds, leila [21:32:53] coming [21:32:54] halfak, I think you killed toby. [21:33:05] lies [23:27:22] Hey, Research & Data Team! Do we need to resched 1/20 staff meeting due to travel? [23:28:02] good point. probably? [23:28:30] K...and will any of you be going to MWDS? [23:29:36] I don't know what that is, so no [23:29:49] oh, the dev summit? nope. halfak, you're The Boy Most Likely To? [23:30:19] Yup. I'll be going to the MWDS [23:30:31] new acronym to me too [23:30:39] K...will resched next 2 staff meetings :) [23:36:12] halfak, I was trying to work out what conference it could be [23:36:31] MediaWiki Distributed Sciences [23:37:10] Monster Warrior Detective's Society [23:37:31] CSCW + DIS == MWDS [23:37:56] Or maybe it's a statement of intend [23:38:04] Where does the M come from? [23:38:06] "what are you doing this weekend?" "Maybe We'll Do Stuff!!!!" [23:38:13] "Merged" [23:38:13] heh [23:38:56] Minnesota Wiki Data Surfing [23:43:49] halfak, I just had a great idea [23:43:57] :) [23:44:03] I'm gonna write a tool in R that automatically issues thanks notifications to people [23:44:11] for any edit that meets our probably-good heuristics [23:44:17] and I'm gonna make people use it [23:44:20] lotsa people [23:44:21] know why? [23:44:31] why? [23:44:45] because I can call it a DISTRIBUTED DENIAL OF SURLINESS ATTACK :d [23:44:49] (I'm gonna have some bomb-ass probably-good heuristics) [23:44:54] heh [23:45:10] ^ that's not fair. I actually laughed out loud. [23:45:24] :) [23:48:53] yay! [23:49:00] seriously, share these heuristics! [23:49:10] I'm going to write it as a WikipediR demonstrator and then hope other people adopt it [23:49:12] * Ironholds nods firmly [23:49:22] ooh [23:49:26] actually I could do it in python [23:49:29] I'll do it in python [23:49:34] Ironholds, will do. This is what my hadoop frustrations are delaying :( [23:49:42] YAY! PYTHON! [23:49:50] I've done things in Python before! [23:49:54] I know. [23:49:57] But more :) [23:50:01] More things. [23:50:38] yes sir! [23:50:50] halfak, this seems like something that would make a really good streaming demonstrator, actually [23:51:02] +1 I absolutely agree. [23:51:06] maybe I build it, with your heuristics. Then we build streaming. Then you build a heuristic-streaming service. [23:51:12] then we add gender filters on to it [23:51:19] gender? [23:51:29] random-ass example because it's what I'm working on [23:51:41] and suddenly we have a series of streams and tools that amount to "thank users for probably-good edits, biasing towards those by new female contributors" [23:51:47] * Ironholds cackles [23:51:56] lol [23:52:04] This is a tiny bit evil [23:52:34] Really, I think this is a good idea and we'll have more luck with a human-computation interface. [23:52:53] yuup [23:53:03] oh, the possibilities. If we could clone yuvi and you and gabriel. [23:53:19] e.g. Log into Snuggle (or something) and it gives you a list of users who should probably be thanked. [23:53:33] If we clone me, I already have a worklist [23:53:42] yuuup [23:53:43] So we're going to need to make a few clones. [23:54:04] I'm down with this [23:54:05] wait [23:54:10] will the clones also make me candied bacon? [23:54:14] because if so I am so down for this. [23:54:25] Presumably. [23:54:34] If they don't that would be a fun paradox to study [23:54:50] Yup!