openstack team mailing list archive

[Ewan Mellor <Ewan.Mellor@xxxxxxxxxxxxx>]

Great stuff Stefano.  These are going to be very valuable metrics.

To start with, I'd draw your "commits per month" graph with a 1 year x-axis.  This is a reasonable baseline for historical stats, and has the particular advantage in this case that it trims off the start-up phase of the project, so the average should go way up.

Cheers,

Ewan.

> -----Original Message-----
> From: openstack-bounces+ewan.mellor=citrix.com@xxxxxxxxxxxxxxxxxxx
> [mailto:openstack-bounces+ewan.mellor=citrix.com@xxxxxxxxxxxxxxxxxxx]
> On Behalf Of Stefano Maffulli
> Sent: 23 September 2011 17:34
> To: openstack@xxxxxxxxxxxxxxxxxxx
> Subject: Re: [Openstack] Community metrics, developers' engagement
> 
> 2011/9/8 Stefano Maffulli <smaffulli@xxxxxxxxx>:
> > in the next weeks I would like to start collecting some relevant data
> > about the OpenStack community. One of the reasons for this effort is
> > to understand it and communicate it to the outsiders better.
> 
> I've spent a few cycles putting together some more thoughts and tools
> to measure the engagement of contributors in OpenStack community. The
> grand plan is to build a beautiful dashboard using business
> intelligence tools, like Pentaho or something similar. I've started
> putting down thoughts and ideas about the datawarehouse on
> http://etherpad.openstack.org/CommunityMetrics.
> 
> Before we get there we'll need raw data, so I've started playing around
> with the tools provided by the folks at FLOSSMetrics
> https://projects.libresoft.es/, especially mailingliststats and
> CVSanaly2. CVSanaly2 is a tool that extracts information out of source
> code repository logs and stores it into a database. It works well with
> git, so I fooled around with it on the nova, swift and glance
> repositories (see the attached plot, as an example). I'll publish the
> queries I used on the wiki early next wee. I'm planning also to play
> with gitdm to see the difference.
> 
> I haven't spent too much brain power on the analysis of discussions on
> mailing list because the mlstats tool expects the archives to be in a
> form that we don't have at the moment. I'll do this later, or if
> somebody wants to help me grab me on IRC (I'm reed there).
> 
> Hope you like this and makes you want to tell me what you want to see
> tracked and measured.
> 
> cheers,
> stef