[OSM-dev] Using changesets-latest.osm

Toby Murray toby.murray at gmail.com
Mon Jan 30 16:14:03 GMT 2012


On Mon, Jan 30, 2012 at 9:47 AM, Josh Doe <josh at joshdoe.com> wrote:
> On Sat, Jan 28, 2012 at 5:01 PM, Toby Murray <toby.murray at gmail.com> wrote:
>> I present ChangesetMD:
>>
>> https://github.com/ToeBee/ChangesetMD
>
> What kind of stats are you (and others) thinking about generating? I'm
> interested in seeing editor stats such as a time series of editor
> usage, which editors people use (predominantly one or a mix, progress
> from using one to another, etc.), the frequency and "quality" of
> comments by editor, what tags people use besides the standard
> comment/source/etc.
> -Josh

My initial use of this data was to get a specific list of changesets
for dealing with the license change. But I certainly had in mind that
it would be useful for some interesting analysis like what you have in
mind.

Off the top, I can say there are just over 700 unique key values used
in tagging changesets. Some of them are obvious errors, some of them
are actually interesting. There are 630,000 changesets with no changes
in them. Didn't P1 create a changeset as soon as it was fired up, even
if no changes were made? There are 239 changesets with 50,0001
objects. Off-by-one error in the API? :)

And that's just the first few things that came to mind. I haven't had
much time to get into the data yet. Heck, I'm still working on
processing my survey data from Christmas :/

Toby



More information about the dev mailing list