[Popcon-developers] Raw popcon data repo schema and access questions

[Bill Allombert]

On Thu, Sep 18, 2014 at 09:01:55AM -0400, Pavan Gupta wrote:
> Hi popcon team,
> 
> While I was pouring over debian package information the other day, I
> happened to wonder what packages were increasingly popular clustered around
> some key packages that I had just installed.  So, I started pouring over
> published popcon data.  It's great data and you all rock for making the
> survey happen, but I couldn't find the level of detail I needed to answer
> my surprisingly complex question.  And that started me wondering what the
> actual central data repository popcon uses looks like and whether it
> contained enough detail to write out a few reasonably simple package
> recommendation tools.

Hello Pavan,
did you read the popcon FAQ ?
(/usr/share/doc/popularity-contest/FAQ.gz)
There are also slides:
<http://popcon.debian.org/paris2014.pdf>

> 1. Is the popcon data collection schema published?

No idea what that means.

> 2. Does popcon collect anonymized, but linked time-series data?

Yes, but only the latest report is kept.

> 3. Does popcon maintain the logs that were generated out of the HTTP posts?

popcon does not. Maybe the Debian sysadmins do.

> 4. Does popcon maintain the emails received from popcon survey participants?

No more than 24 hours (for rollback in case of server errors).

> 5. Does popcon publish its server-side code for generating final tabulated
> results?

Yes, in /usr/share/doc/popularity-contest/.

> 6. Would it be possible to receive access to the raw popcon data?

Sorry, this is restricted to Debian developers with an account on
popcon.debian.org.

> And I know this is where the lines blur on privacy.  Up front, I should be
> clear that I have no monetary interest in this, I have no interest in
> identifying anybody or any machine, and anything and everything I would
> write would always be MIT licensed -- heck, you can just have it without
> attribution if you'd like it.  Maybe most importantly, I'd also be happy to
> allow you to gate the results that come from any tools/analysis/etc that
> might come out of my questions below:

Sorry, I would not have time to follow through this process.

Cheers,
-- 
Bill. <ballombe at debian.org>

Imagine a large red swirl here.