Source Data

Converting data from a variety of formats, usually ERSI Shapefiles although I am seeing more data in GeoJSON format, is a time-consuming process. Since there also seems to be a trend towards government sources moving towards an ArcGIS solution, which often doesn't support downloading. At the same time, more data has become available under an unencumbered public domain license. I download what I can. Please use upstream sources for the original data files, mine are mostly here as part of documenting the import process, or if the original cease to be available.

I've also done the hard part of converting the data, and chopping up the huge files into more managable pieces in the hope others can use this to correct existing data, or import what is lacking into OpenStreetMap. As all the conversion has been initially validated, importing it will be much less time consuming.

Before importing anything, you should read the OSM Import/Guidelines. I've used normal OSM tagging usage, and when in doubt, have used the OSM Tagging List and the proper tagging. For many areas this data often exists in OSM, but lacks metadata. So these souce files can be used to improve metadata in OSM. The richer the metadata, the better the final maps are. Especially when mapping an areas you're familiar with, you can more easily validate the metadata.

For anything more than a trivial change here and there, you should strongly consider following the process as outlined above. Bulk imports need to be discussed with the local OSM community, if there is one. For rural areas, there probably isn't one, although some traveling recreationalists also OSM contributors. People may care about an area but not live there. You can use this documentation on the process I use, with more here on the specifics of importing this data.

Since my primary focus is maps for emergency response and recreation, the data we like intially is what kind of road or trail is it ? Can I bring a fire truck, or only a brush truck, or only a UTV ? What is the surface, the width, the roughness, etc... Old obscure 4WD roads are very useful for both wildland fires and backcountry reescues. We also use house addresses, as most dispatch messages are to an address, not GPS coordinates. Building footprints are useful too, as accidents on large ranches often seem to happen in remopte outbuildings.

I have found many bugs buried in the original datasets, and road conditions change due to erosion or heavy usage. Or your favorite camp site may now have no tresspassing signs everywhere. I also do a lot of ground-truthing of the data, because to really know you have to see it with your own eyes. I have a series of tech articles that document ground-truthing map data.

Highway Data

I've been cleaning up a lot old crufty data I had on this site and updating everything. The data is currently focused on remote highways, as I plan to fix the TIGER import mess, as it keeps leading to major navigation errors. I have the data broken down by State. All the external sets I'm using for conflation have been processed to be ready for conflation using my OSM Merge project.. I then converted the ERSI Shapefiles to GeoJson using ogr2ogr. For the conversion to OSM syntax, and the deletion of many data fields that aren't needed, I use some simple utilities I wrote. Then all the names were converted to OSM standard, abbreviations expanded, etc... Then the data was cleaned up by deleting all the orphan nodes that don't actually represent anything, or features that lack a geometry. You can of course compare my conversion by looking at the original data file. There is detailed information on the actual conversion here. It's public domain licensed, and has been used before for imports.

At that point the files are still huge and hard to work with. To make my life easier, and encourage other to help with this project, I've chopped up the large files by various boundaries. If you import any of this data from these files, don't forget to simplify the ways before uploading to OpenStreetMap.

Top of project