[OSM-dev] TIGER upload automation

[Dave Hansen]

On Thu, 2007-08-30 at 19:13 +0100, Robert (Jamie) Munro wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Dave Hansen wrote:
> > This is the actual queue that I'm using to drive the upload automation.
> > If you want your county done sooner, let me know, and I'll pop it up in
> > the queue.
> > 
> > 	http://www.sr71.net/~dave/osm/tiger/upload-queue.txt
> 
> How did you decide initial order for that queue?

1. My state first :)
2. Somebody asked for counties in MD and WA
3. Alphabetically

> I notice that the population of US counties varies from nearly
> 10,000,000 (Los Angeles County, CA) to just 60 (Loving County, TX. And
> yes, that is sixty, not six hundred or sixty thousand or anything.
> 
> See http://www.census.gov/popest/datasets.html#county for the full dataset.
> 
> It might be sensible to start with the biggest counties on that list and
> work your way down, so that we get the maximum percentage population
> covered for the lowest possible upload time at the beginning. It looks
> like about half the population lives in the largest 150 counties, which
> at 6 counties/day will take less than a month.
> 
> Of course, it's probably the case that those 150 counties also have most
> of the TIGER data, so it might be sensible to divide the populations of
> each county by the data size and order by that figure.

First of all, I like grouping it by state because I think a large areas
with complete coverage are more desirable than a couple of large
counties dotted across the US.

Other than that, I think the top priority simply goes to people asking
for their counties to be uploaded.  I'm not really compelled to change
the queue around otherwise, just because of the reasons you stated
above: large counties cover more people, but they also take more time.

If somebody wants to put together a correlation between number of OSM
objects in a county and its population and prioritize counties with a
low number of OSM objects per capita, I might be more compelled.

Here's a set of nodes/segments/ways broken out by county and by state
for ever TIGER file I've generated so far:

	http://dev.openstreetmap.org/~daveh/county_osm_stats.tar.gz

-- Dave