In Canada, we select IDs that (1) are publicly available (even if not easily discoverable) and (2) are least likely to change over time. Using names for identifiers is generally a bad idea in Canada since electoral district names are changed all the time. Just this year, Parliament changed about 10% of federal electoral district names. Note that in Canada, you can change a district’s name (and even its boundary) without formally abolishing the district - so the ID should be something that persists through name changes.
In BC, they don’t use numbers, but they do use three-letter codes. See here:
http://www.elections.bc.ca/index.php/maps/electoral-maps/geographic-information-system-spatial-data-files-2012/ Their geospatial files do not contain any numerical codes, just the three letters.
Saskatchewan actually does have numeric codes, but they only appear in one dataset, which was previously annoying to work with. I’ll update the identifiers to use those numeric codes, as they are more future-proof.
As for other provinces, Quebec has non-consecutive numeric identifiers, which is strange, but they use those identifiers in lots of places, so we trust them. Manitoba and Newfoundland and Labrador simply don’t have any publicly available identifiers, so we use the names.
There is, sadly, no consistency across provinces, and our data reflects that. We aren’t in a position to impose consistency on the provinces.
Cheers,
James