Re: postal code-->electoral district: a free + legal way
Posted by
Tracey P. Lauriault on
URL: http://civicaccess.416.s1.nabble.com/postal-code-electoral-district-a-free-legal-way-tp1518p1531.html
Been away for a couple of days!
These folks did some interesting work here with postal codes but I do not know the level of accuracy -
http://www.voteforenvironment.ca/
Jennifer, if you go through the archives of the
civicaccess.ca list you will find two or three threads that discuss this file and its accuracy & access issues.
Russel over at
Copyright Canada has developed a tool here -
http://www.digital-copyright.ca/edid/postal, and has an MP letter writing tool created to request for access to this file that might be helpful for the discussion thread you forwarded earlier.
Letter re-postal code file and the tool he used to help you generate it are here -
http://www.digital-copyright.ca/edid/letter5.
I also know that
Mike over at
Open Concept (
http://openconcept.ca/services) has done some work with the postal code file and perhaps he can point us to some of the issues he encountered.
Daniel is correct, the issue here is accuracy, you want to make sure that you continuously have the updated file and that you are working with polygon to polygon to ensure you have the correct postal code in the correct riding. The other tools are not accompanied with metadata and therefore it is uncertain if they did a quality check or what base data they used. This is a typical problem with may great hacks, no metadata for the user to evaluate the tool, assess the algorithm nor trace the data source.
Cheers
Tracey
On Thu, Jan 29, 2009 at 12:02 PM, Joe Murray
<[hidden email]> wrote:
The Elections Canada / StatsCan list for PC -> EDID is known to have
significant data quality issues unrelated to a postal code geographically
overlapping riding boundaries. There are up to 7 ridings for a single postal
code in their data last I looked, with tens of thousands of postal codes
having multiple entries if I recall correctly. The methodology they've used
to create the table is the issue.
> <
[hidden email]> wrote:
> >
> > Isn't that a problem with all postal code to electoral district
> resolvers?
> >
> > Jennifer
>
> Well, something is definitely fubared with these lists, because they
> don't all return the same results.
>
> Technically though, it should be an area overlap. For those postal
> codes that overlap 2 or more EDIDs, all possibilities should be
> returned in a query.
>
> Otherwise, we mislead a fraction of voters.
>
--
Tracey P. Lauriault
613-234-2805
https://gcrc.carleton.ca/confluence/display/GCRCWEB/Lauriault