Durham County, North Carolina Building Import
To add missing buildings in Durham County to OpenStreetMap without creating duplicates.
The import will be conducted in small pieces of about 500-1000 buildings from Durham Open Data. This will allow me to check them all individually and add building type (house, detached, apartment, shed, etc..). The import started on June 28 2018 and will continue until sometime in August.
Data source site: Durham Open Data
Data license: Open Database License
Type of license: Open Database License (ODbL)
Link to permission: N/A because of ODbL.
ODbL Compliance verified: Yes. I asked the U.S. slack OSM group. (osmus.slack.com).
OSM Data Files
I will convert the files from KML to OSM as I import each small chunk. I will add the tags "building"=* and "source"="Durham Open Data".
Many small imports over the course of a couple months that will add almost every missing building. The import will include tons of manual work to check and ensure the quality of the data being imported. The import will be completed via JOSM. Durham Open Data has about 103549 buildings.
Data Reduction & Simplification
I will use "building"=[whatever] and "source"="Durham Open Data".
I will use "source"="Durham Open Data", "source:website"="https://opendurham.nc.gov/explore/dataset/buildings/export/", and provide adequate comments.
Very little data transformation is needed except for the manual building quality checking. I download the buildings as a KML file and open them using the JOSM opendata plugin.
Data Merge Workflow
I will be working on this import alone. It is very simple, because it only involves adding ways with simple tagging to the database. No merging or other complicated nonsense is required.
For each of the about 100 smaller chunks of buildings:
- Download from Durham Open Data as KML.
- Open with JOSM via opendata plugin.
- Add building type and remove/fix bad data.
I will be using the account LeifRasmussen_import to do the import. This should make reverting changes easier in the unlikely event of a screw-up.
I will manually conflate the data to ensure the best data quality.