[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [Debconf-team] help needed for penta datamining

Steve Langasek dijo [Mon, Mar 24, 2014 at 11:32:34PM -0700]:
> Hi folks,
> In the process of trying to finalize the budget, I've arrived at the point
> where most of the uncertainty relates to how many people will be
> attending/sponsored/registering as paid attendees.  While we won't really
> have final numbers on this until the conference itself, it would be useful
> for forecasting to get a sense of how these numbers have varied (or not)
> from year to year.
> (...)
>  - how many total people attended each DebConf, by day?[1]
>  - how many people stayed in sponsored accomodation at each DC, by day?[1]
>  - how many people attended each DebConf in each of the paid registration
>    categories (professional/corporate)?
>  - how many people received travel sponsorship for each DC?
>  - how many people were registered to attend each DC?

Hi Steve,

I think I can put some time to this tomorrow afternoon-ish; I will be
mostly offline, but I'll try to do this, at least partially.

If not tomorrow, by Thursday I think I can look at this a bit more
quietly back in my office.

> Best would be to have this data made available as a standing report that we
> could reference for the future as well.  FWIW, there used to be just such
> data available at <http://debconf-data.alioth.debian.org/stats/>, but this
> disappeared following the last alioth crash. :/  Other sources of historical
> data include <http://penta.debconf.org/stats> and
> <http://rkd.zgib.net/http/debconf/> - it would be keen if someone could
> reconstitute these reports and get them re-published somewhere suitable. 
> Or if someone had a cronjob where these were being generated, and has a good
> way to set it up again, maybe that would be even better!

I can make a script for this, that can be run periodically as a cron
job. Don't want to promise too much, as my free time is somewhat
limited, but it should not be too hard.

Reply to: