[GIS] Using CanVec from Natural Resources Canada

canadadata

It has now been a little over 3 years since Natural Resources Canada released CanVec, a national coast to coast to coast spatial dataset at 1:50,000 or better nominal scale. In the interverning time there have been several updates, we now sit at Edition 6, and there are scheduled updates every 6 months, indefinitely.

Canvec replaces the National Topographic Database (NTDB) which was both expensive, ~$240 per NTS tile if memory serves, and definitely not libre. With the NTDB we had to pay royalties for every map that went to another party, even if we gave it away. The royalty was reasonable, on the order of 25 cents each, but we had to pay in advance, in $5,000 chunks, with a minimum of $8,000.

Canvec is a well modelled internally consistent dataset of good quality, better than the NTDB, and it's FREE, both libre and gratis. What puzzles me, greatly, is the virtual silence online* about this fantastic resource. There is little chatter about Canvec. Perhaps this scale of data just not that relevant south of 60°N and you're using something different. Perhaps the naming conventions are so opaque that people just can't drum up the energy to figuring out what's what.

Where are the resources and tools to make using this stuff easier (or even possible)?

Best Answer

I use the CanVec offerings somewhat regularly in building teaching assignments for profs, but the data require a good amount of work to go from what's provided to what a novice GIS user (who may or may not actually be interested in GIS) can handle.

As an example, if I want to supply a layer of fire station points in our region as part of an assignment, the process is as follows:

  • Download and extract the zip file for each of the 4 tiles that intersect our region (not CanVec's fault, the boundaries between tiles have to go somewhere)
  • Look up in the documentation what the number is for the buildings and structures layer, and what the code is for fire stations
  • Isolate the 8 shapefiles that have the information I need (1 polygon and 1 point layer for each tile)
  • Merge the 4 polygon layers and the 4 point layers
  • Select and export the features I need in each
  • Convert the selected polygons to points
  • Merge the two point layers I now have
  • Remove the duplicated points, if any (there are sometimes features in the polygon and point layers that represent the same feature on the ground, this tends to bug me more than anything else)

That's a decent amount of processing to get what I need, but it's not hard and probably not more than what the average GIS professional has to do to wrangle their data.

That said, maybe the complexity of the data model scares people off (particularly the naming conventions, as the question mentioned). I think you can usually find what you're looking for in CanVec, as long as you know how to look and are willing to read the documentation.

The other (more anecdotal) thing is that in my experience, compared to NTDB, there are a lot more features categorized as "unknown" in CanVec. This sometimes includes features that were properly categorized in NTDB, but then became unknown in CanVec.

Related Question