[Analytics] Wikipedia page views

2014-03-24 Thread Burton DeWilde
Dear Toby, I recently saw your comment on a blog postby Magnus Manske regarding the lack of Wikipedia page view data besides the oft-overloaded http://stats.grok.se/. I was wondering if there's been any progress at WMF on building a more stable, central, an

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Alex Druk
Hi Burton, We just opened a new site www.wikipediatrends.com that show Wikipedia page view data. Our site is very similar to existing http://tools.wmflabs.org/wikiviewstats/ and http://stats.grok.se/, but use slightly different approach to calculating and presenting data as well as allow compariso

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Magnus Manske
Quick question: Does this have en.wp data only, or can I query (as in CSV) other wikipedias/projects? And, can I limit the data range (not really necessary, but less data to transmit)? On Tue, Mar 25, 2014 at 9:09 AM, Alex Druk wrote: > Hi Burton, > > We just opened a new site www.wikipediatren

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Alex Druk
Hi Magnus, Only en.wp for now. We will wait and see how popular it is before adding other projects. You cannot limit data range in csv now, but the size of the response is usually < 10 KB. By the way, many thanks for your great work! Regards, Alex On Tue, Mar 25, 2014 at 10:53 AM, Magnus Mansk

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Hay (Husky)
@Alex: that's really awesome. Thanks for providing a stats.grok.se alternative. Really looking forward to other languages as well, and maybe throw in Commons in the mix as well? -- Hay On Tue, Mar 25, 2014 at 11:06 AM, Alex Druk wrote: > Hi Magnus, > > Only en.wp for now. We will wait and see ho

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Alex Druk
@Hay: Thank you. The site is still in very early stage of development. We would like to get constructive criticism from the wikipedians from this list first. Our resources are very limited and we cannot include all other wiki projects now. What languages would you like to see first? What projects?

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Federico Leva (Nemo)
Just saw an update from Henrik: «hopefully within the next two or three weeks the capacity of stats.grok.se will be quadrupled». https://en.wikipedia.org/w/index.php?title=User_talk:Henrik&diff=600917917&oldid=600897425 Alex Druk, 25/03/2014 10:09: We just opened a new site www.wikipediatrends.

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Alex Druk
@Nemo: Ooops! Will be fix soon. yes, we have issue tracker at https://github.com/sergeychernyshev/wikitrends/issues?direction=desc&labels=bug You can also submit any at http://www.wikipediatrends.com/ContactUs.php On Tue, Mar 25, 2014 at 12:30 PM, Federico Leva (Nemo) wrote: > Just saw an updat

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Hay (Husky)
On Tue, Mar 25, 2014 at 12:40 PM, Alex Druk wrote: > @Nemo: Ooops! Will be fix soon. > yes, we have issue tracker at > https://github.com/sergeychernyshev/wikitrends/issues?direction=desc&labels=bug > You can also submit any at http://www.wikipediatrends.com/ContactUs.php The Github page gives a 4

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Dario Taraborelli
Hi Burton, nicely done (and yay for using dygraphs) – with what frequenty do you expect wikipediatrends to ingest new data from the raw pageview dumps? I assume it’s once a month? Dario On Mar 25, 2014, at 2:09 AM, Alex Druk wrote: > Hi Burton, > > We just opened a new site www.wikipediatr

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Dario Taraborelli
On Mar 25, 2014, at 7:01 AM, Alex Druk wrote: > @Dario: thanks. yes, we renew the site once in a month, usually around 10th > of each month because dependence on dumps. > And yes, we plan to introduce JSON awesome also, I noticed some inconsistency in the heading/titles that you may want to

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Alex Druk
@Dario: thanks. yes, we renew the site once in a month, usually around 10th of each month because dependence on dumps. And yes, we plan to introduce JSON Alex On Tue, Mar 25, 2014 at 2:32 PM, Dario Taraborelli < dtarabore...@wikimedia.org> wrote: > apologies, s/Burton/Alex :) > > one more ques

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Jane Darnell
Fascinating website, and I love the comparison option - I just compared page hits on Haarlem vs Leiden and I guess the spikes due to tourist attractions. 2014-03-25 12:48 GMT+01:00, Hay (Husky) : > On Tue, Mar 25, 2014 at 12:40 PM, Alex Druk wrote: >> @Nemo: Ooops! Will be fix soon. >> yes, we ha

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Dario Taraborelli
I haven’t checked the raw logs to compare them with the visualization but I think we should QA the data: the raw (unsmoothed) series for Eros shows a spike on 2/14 (Valentine’s Day, predictably) with 6,920 pageviews, while stats.grok.se reports for the same date 3,209 page views. I don’t think a

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Alex Druk
@Dario: I do not have time to check original files now, but I believe that difference reflects that we show aggregated data (i.e. data for the article PLUS all it's redirects). However, I would check raw files also. On Tue, Mar 25, 2014 at 3:51 PM, Dario Taraborelli < dtarabore...@wikimedia.org>

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Dario Taraborelli
apologies, s/Burton/Alex :) one more question: is there any plan to add a JSON interface on top of the CSV download? Many people have relied on stats.grok.se JSON output for years and it would be fantastic to have wikipediatrends return data in the same format. Dario On Mar 25, 2014, at 6:27

Re: [Analytics] Wikipedia page views

2014-03-25 Thread Toby Negrin
Hi Burton -- Thanks for this. I'm glad the Wikipedia data is useful, even if it's difficult to access at this time. As Nemo reported, we're currently working with Henrik to get him a better server and it should be on it's way to him now. We're hopeful that modern hardware and SSDs will really hel