I have started an import process of the EPA dataset.
Here are the tags that I have used in the import :
- landuse = industrial
- man_made = environmental_hazard
- name = from the KML file
- ref = http://iaspub.epa.gov/enviro/national_kml.registry_html?p_registry_id=110010106081
- source = http://www.epa.gov/enviro/geo_data.html
There are a number of issues with the EPA records :
- A large number of the points are not accurate. They need to be moved or deleted. This is one good reason to have it publicly reviewed by OSM.
- The symbols in my import are all the same. There are many different types of EPA records.
- The records that have been deactivated, to get this data, it is needed to check the EPA database. This will be a long process.
- The names have to be renamed to normal case, I found a Perl modules to do that. Some names are abbreviations.
- There are existing nodes that describe some of these features and duplicates need to be removed.
- The current Mapnik rendering rules display these nodes in large text at zoom level 15 and above. This will encourage users to edit the nodes as they see them, further muddying any data cleanup efforts.
- The longer the data sits in the map, the more difficult it will be to develop a cleanup algorithm.
Here are a set of sites that have been deleted, should be removed or checked. http://www.epa.gov/superfund/sites/query/queryhtm/npldel1.htm
Here is some ongoing work to update/correct the imported data. https://code.launchpad.net/~jamesmikedupont/+junk/EPANatReg