[Wikimediaindia-l] You will love Erik Zachte for this :)

[Erik Zachte]

Anirudh:
> Hopefully, the under-reported page views issue will be resolved soon!

I rcvd notice server overload has been contained.
It will take some time to patch data from last weeks.
Underreporting is much less severe than in previous incident, maybe 5% for
November 2010.

For permanent solution the whole traffic log infrastructure needs overhaul.
Different software will allow distribution of load over several servers.
Nimish Gautam is working on this, but this is a major change and may takes
several months.

------

Ravi:
> http://bit.ly/gdH21O

Thanks. I very much like this reach metric, should make it a wikistats table
for all languages some day. 

------

Ravi:
> Do we have such region wide stats for other Wikimedia projects like Wiki
source, Wiktionary also? That will be helpful too.

Shiju Alex: 
> Here is another request from Indian language wikimedians to create region
wise report for sister projects also.

Point taken (the question was actually to be expected). I will need to find
a little time to do this.

------

Sundar:
> Shall we expand the list to include the remaining South Asian languages
under the title "Wikipedia Statistics - South Asia"?

I will wait for consensus to arrive. How many languages would that add to
the list?

Also it could complicate matters some day when number of speakers per
language for the region would be factored in.
Somehow Wikipedia knows secondary language speakers on a global level for
most languages, but per country/region?!
I expect quality of census data will differ per country. Does census even
measure secondary language speakers?

------

Achal:
> is there any way you can easily and comprehensibly integrate India edits
and readership to EN on the India page? 
Gerard:
> Including English on the page for India only makes sense when the traffic
is the traffic generated for the English Wikipedia in India. 
Achal:
> Yes Gerard, I assumed it was obvious that we're talking about India-based
traffic and edits to EN

I added EN today, which makes this report consistent with other region
reports. 

It would indeed be nice to extend or modify or provide as alternate view
following reports,
based on view and edit counts solely from the region,
http://stats.wikimedia.org/EN_India/TablesPageViewsMonthly.htm
http://stats.wikimedia.org/EN_India/Sitemap.htm

Doing this for full report would require significant update in data
collection and presentation layer, 
for which resources are currently not available.

I could start with an extra row or column (depending on the report), for
share of views/edits from the region.

That kind of info is already collected and used on 
http://stats.wikimedia.org/wikimedia/squids/SquidReportPageViewsPerCountryBr
eakdown.htm
http://stats.wikimedia.org/wikimedia/squids/SquidReportPageViewsPerLanguageB
reakdown.htm

But even this will take a while, for three reasons:

1 Other priorities 
2 Both data streams (per country breakdown of views and edits, monthly page
views per wiki) are entirely different processes. 
Different data sources, different servers, different accuracy, different
periodicity of jobs. One process is fully automated, 
the other not quite yet. 
3 While current sampling rate of squid logs is quite satisfactory for page
views, 
it is really too low for accurate breakdown of edits per country per
language.
This is under consideration and will be improved. 

Erik Zachte