[Strategic] OpenStreetMap Growth Rate

[Kai Krueger]

 > Hi,
 >
 > Oliver Kühn wrote:
 >> can anybody organize the OpenStreetMap growth rate per region? My
 >> idea was to have an indicator like [new nodes] / [100 square
 >> kilometers] over the last x month for different countries or regions?
 >
 > I think this is of interest to a wider audience and should be discussed
 > on talk; Kai Krueger has done some preliminary work in that direction
 > and I could imagine he'd be able to provide interesting stats if briefed.

There are quite a few set of statistics of all sorts and kinds out there 
already. Perhaps the closest to what Oliver was looking for that I know 
of is http://stat.latlon.org/ by Komяpa. Other useful statistics are 
completeness measures (comparing to official lists) like the street 
lists in Germany ( http://osm.gt.owl.de/Strassenliste/ ) and the 
comparisons to OS Locator in the UK ( 
http://humanleg.org.uk/code/oslmusicalchairs/, 
http://dl.dropbox.com/u/685499/OSM_OSLocator_stats.csv ), the usual OSM 
statistics and many more.

Statistics are nice, but don't forget the moto of "Never trust a 
statistic that you haven't faked yourself" so it is important to have a 
clear objective of what you want to show and for what purpose before 
creating more.

Is it for marketing purposes, i.e. show wonderful growth curves all 
pointing to the sky, or is it to identify problems of a specific form? 
So Mikel's idea of gathering some ideas and define what we want to show 
and why on the wiki sounds like a sensible approach and I am sure then 
some of us will be able to create the necessary graphs to go with it.


Back to the specific statistic of nodes / area broken down by country. I 
am not sure if that is necessarily the best metric, as there are a 
couple of confounding factors in there. Perhaps the biggest is the 
inherent density of the country[1]. A country like Monaco, Macau or 
Gibraltar are probably going to have higher node / area counts than e.g. 
Mongolia, Greenland or Australia. The other big confounding factor would 
be imports. Although imports may be good to kick start an area (there is 
considerable debate about that as well...), in the end it is the 
community that matters, as only with a strong community can we achieve 
the high quality and up to datenes that OSM aims for. So perhaps 
something like users / inhabitants may be more reasonable. But again, it 
heavily depends on what you want to show.

Kai

[1] 
http://en.wikipedia.org/wiki/List_of_countries_and_dependencies_by_population_density

 >
 > Bye
 > Frederik