[OSM-talk] US City Import?

[Beej Jorgensen]

Beej Jorgensen wrote:
> On a whim, I wrote a python script that takes US Census data and USGS 
> GNIS place name data, and produces an OSM file of points with hamlets, 
> villages, towns, and cities.

Found an issue after uploading a little test area around Berkeley, CA 
(the uploaded city name nodes have since been removed.)

There are 15,000 duplicate place names in the GNIS data.  For example, 
there is a Thousand Oaks, CA in Ventura County, population 124,000, and 
a neighborhood by the same name in Alameda County, CA, population about 
1,000.  The Census data doesn't differentiate on anything other than 
name.  The GNIS lists them both as "Populated Place".

There's not enough information in either data set to go on.  I'll need 
to find some better data, or it's going to be 15,000 edits to get the 
city sizes to show up correctly. :(

-Beej