Espírito Santo (state)/IJSN road import
Instroduction
IJSN or Instituto Jones dos Santos Neves is the state institute of statistics and cartography for the Brazilian state of Espírito Santo. The holds a public archive with several shapefiles containing data that are missing, or partly included in the OpenStreetMap databank. The purpose of this import is to compare two shapefiles containing road infrastructure, with the data in OSM, and supplement the missing datas
License
The license have been confirmed by contact with IJSN as Public Domain, no attribution necessary, but for the sake of traceability, the source=IJSN will be included in changesets.
Confirmation about License (PD) posted on forum
Quality of data
Overlaying data from IJSN onto OpenStreetMap shows that in general the geographic shape of data in OSM is close to complete, with similarities enough to use python shapely for object comparison.
What to import
Both shapefiles contains name=*s, number of lanes=* and surface=*, the name tags are not directly suitable in OSM as they are presented in the shapefiles, but enough information is available for a python script manipulating the shapefile can correct the names (remove abbreviations, extract ref numbers, etc) so we have at hand name=* for almost all road segments in the data set (the ones that don't have are assigned as noname=yes), a few with a valid alt_name=*, and many of the major highways with a valid ref=* in accordance with Brazilian tagging standard. There are also identified a few bridge=yes and bridge=viaduct as well as junction=*
Where conflicting information are presented between the source data and the imported data, an individual error list of objects are generated for human investigation. Imported data are from 6 to 2 years old, if data already in OSM are based on user survey of newer date than imported data, we want to keep the correct one.
Before import status:
The script is made to produce some statistic output. Dates to runs, with output will be posted here.
The scripts
Source code for the entire import are available on Github.
Schedule
It takes 8 - 10 hours to run the entire analyse, divided between the 78 municipalities. Each municipality will be analysed and uploaded in sequence. Date to be scheduled after completing discussion with community and fine tuning of scripts.
Flares
Links to error files will be placed here after run of the import. flare_example.json