Geographic Placename Data

Where does DMoz gets its placename data from for the regional categories?

Our non-profit http://e-democracy.org is considering the use of local ODP listing data for a "town square" like presentation of links to local government, media, libraries, schools, etc..

We are also thinking about the creation of a simple public domain database of local government "portal" web sites and would like to leverage/connect it with ODP where possible. If we use government generated placename databases (e.g. http://geonames.usgs.gov/gnishome.html for the U.S., http://earth-info.nima.mil/gns/html/ for intl places) how well/deep will it match ODP data?

Steven Clift

I can share my experience as a Regional Editor in Arizona: I get my placename data from a variety of sources, including the Arizona Blue Book and personal knowledge of the state. Granted, much of the data was in place when I began editing in Arizona, but as a high-growth state, I probably add about one new location a month.

Others, who have been around longer, can probably give a better answer.

I think you are looking for an automated way to get those data. That is not what we do: We rely on our editors to build the structure, one category at a time.

Early on, there was a bulk creation of US locality categories. I'm not sure where staff got the list, but it may have come from the USPS. Although a large percentage of the localities were really places, some actual places were missed, and some places may at one point in history been significant but now are barely a crossroads with a couple houses. The Regional editors constantly refine the localities, based on state and local definitions (or lack there of) of cities, towns, municipalities, townships, villages, and whatnot. They also refine the list based on whether or not there are sites for the category - deleting empty locality subcats when no sites can be found, and adding them when there are sites and the place is distinct enough (regardless of its legal or USPS status.)

At this point, I would say that the ODP list of placenames is pretty unique.

<added>
I didn't look in depth at the government placenames site you linked to, but I have looked at similar things. I would caution against starting with something like that, because it will have far too much detail. They tend to have every hill, valley, stream, and hamlet that was ever recorded in the last 250 years in those, which just doesn't correspond to today's reality.

The US Census bureau has a list which groups its "official" placenames into geographical metropolitan regions. This might be a closer match to the ODP, but again, there will be a lot of variation.