[Talk-ca] Talk-ca: Bulk Import of Address Range in GTHA from Metrolinx, Second attemps
jfd553 at hotmail.com
Tue Feb 9 14:24:13 UTC 2016
Here are few comments/questions I wish to add to Stewart comments.
1 - Stewart is right about questioning the StatsCan license...
Why don’t you use GeoBase data instead? As far as I remember, their license is compatible with OSM and the Canvec road network (streets and address ranges) was generated directly from GeoBase data!
2 - Stewart is also right about government data accuracy. You should never assume that their data (even from GeoBase) are better than OSM. However, I understand from previous emails that you are validating roads location by using Bing/other imagery, which is good! Just make sure imagery fits on available GPS tracks (especially if there are many of them).
3- Made properly, it will benefit to the community.
From: Stewart C. Russell [mailto:scruss at gmail.com]
Sent: February-09-16 07:16
To: talk-ca at openstreetmap.org
Subject: Re: [Talk-ca] Talk-ca: Bulk Import of Address Range in GTHA from Metrolinx, Second attemps
Thanks for sticking with us! And also for including the sample data as an attachment.
One observation: Statistics Canada does not use the Open Government Canada Licence (or even a derivation of it), but its own:
Although OSM has used previous road network information before, it appears (to me) that this StatsCan licence is different from the one reviewed for the 2009 import. It has some prescriptive terms that may make its use incompatible with OSM. For instance:
[you shall] not merge or link the Information with any other
databases for the purpose of attempting to identify an
individual person, business or organization
I read this that you can't use StatsCan data as part of a larger geocoding database, which is one of OSM's uses.
I see that the Triplinx source has been added to the OSM Contributors page, too.
> At a high level, our process of identifying gaps in the address ranges
> is summarized below:
> · For each side of each StatCan road segment with a valid
> address range (start value and end value exist and are different):
Can you confirm, please: you're checking StatsCan address ranges against
**OSM** street segments? StatsCan (and any government data) cannot automatically be assumed to be more correct than OSM data, and we sometimes have to adjust imported data to match air photos or foot surveys.
For consistency, address range segments need to be parallel to OSM street segments.
In addition, please consider:
* not adding very short address ranges. There were a few in my neighbourhood that appeared to be only two houses long. Address points would be better for those.
* that every set of address range end nodes has an address range way associated with it. Again, there were several sets of range nodes with no range ways associated with them near me.
> *Benefit To The Community:*
> … From our perspective, adding address data for areas/streets that
> don’t have this data is a step in the right direction.
I'd absolutely agree, and apologize if I appeared to suggest otherwise.
Talk-ca mailing list
Talk-ca at openstreetmap.org
More information about the Talk-ca