GeoBase Import

From OpenStreetMap Wiki
Jump to navigation Jump to search

The GeoBase dataset covering all roads in Canada can now be imported into the OpenStreetMap database! Development of scripts for conversion and import will take some time. This page should be used to coordinate the process and also provide updates on how things are progressing. There will also be a lot of discussion on the talk-ca mailing list.

Prelude

We are all part of the OpenStreetMap community. The community expects automated edits and imports do follow community guidelines.

One of those guidelines is proper attribution. The imports must be attributed here.

For the current status, see Canada Import Status.

GeoBase WMS Service

GeoBase WMS Service is available. As on the talk-ca list for JOSM link.

Agreement Announcement (18 November 2008)

See Announcement.

Import Process

Each dataset will require a different import process. See the data set specific pages for details.

The general procedure for an import is:

  1. Determine which GeoBase attributes will be imported and how they should map to OSM tags.
  2. Develop or adapt scripts to import the data.
  3. Import some test data and get community feedback.
  4. Import the complete data.
  5. Perform cleanup.

See Common Tags for how to deal with attributes common to multiple GeoBase datasets.

To find the GeoBase NRN complete or excluded file for your area, see the spreadsheet of GeoBase tiles and download the file for your area. Complete includes the entire NRN dataset for a GeoBase tile, excluded includes only roads which were NOT imported because they are very close to existing OSM data.

AutoMatch Process

The automatch process goal is to merge the OSM existing objects with the geobase objects in order to get the best possible result in an automated way. It should add nodes to osm and geobase ways where the two datasets intersect. It should also try to keep the best information of both datasets (hybrid).

Based on local expertise, an optionnal argument should allow to choose the reference datasets. If someone judge that one dataset is of better quality for a giver area, then this dataset should have an higher priority level in the automatch process.

Algorithm

For now, the proposed algorithm is:

For each segment in geobase
 	If an end of the segment forms a junction with another segment
 		If that segment is a OSM segment(see 1)
 			AddNode()
 		endif
 	endif
endfor

where addNode()
adds a node onto the geobase segment that is on the osm segment.

This algorithm was originally used in the workflow using OpenJump with the RoadMatcher plugin. However, there is no clear benefit of using this method, especially in areas which already contain dense OSM coverage. Therefore the current workflow constitutes of converting a portion of the Geobase GML file to OSM, open it in JOSM, copy it into existing OSM data, and upload the result.

FAQs (Frequently Asked Questions)

How long will it take to import?

The import is being done by volunteers on a free time basis. We also want the process to be as accurate as possible and would rather not have to delete and redo portions of the import. The import process started in January of 2009 and is expected to take many months, see the status of each dataset for progress updates. As of November 2009 about half of the Geobase NRN data has been imported.

How can I help?

  • Check out the Geobase import how-to page.
  • Join & read talk-ca.
  • Pick a data set you want to work on.
  • See the pages on each data set to learn what has already been done.
  • For the areas that have already been imported, please consider helping fixing way connectivity and other topology issues where existing OSM data and newly imported data meet (see After the Import section at Import Page).

Can I import other datasets from the Canadian government?

The question came up on whether or not GeoGratis was automatically included with the GeoBase import announcement. GeoGratis is from Natural Resources Canada (Government of Canada), which provides a whole lot of information at no cost to users.

The data must be licensed in an OSM compatible fashion. The GeoBase and CanVec data is, ask for clarification on talk-ca before importing other datasets

Links

  • TIGER – US equivalent
  • LINZ – New Zealand equivalent