[Imports] [Talk-ca] [Import] Ottawa Buildings & Addresses [Statistics Canada project]

Christoph Hormann chris_hormann at gmx.de
Mon Oct 24 22:43:55 UTC 2016

On Monday 24 October 2016, Stewart C. Russell wrote:
> > Could you please clarify what the alleged source of the building
> > data to be imported is?
> Yes, as far as I understand it, at least: these data to be imported
> were given to a group of OSM contributors by the City of Ottawa. They
> have explicit permission to include it in OSM from the City. The only
> place you can inspect the data is on the contributors' own hosting
> sites, as the city doesn't host it anywhere. Details of the licence
> and permissions are on the Ottawa import information page.

I would then suggest to contact the person who provided the files for 
metadata and specifications on those, in particular dates and methods 
of survey, processing applied, especially coordinate system conversions 
and specifications on what exactly is contained in it (i.e. what the 
definition of a building is here and if it's ground footprints or roof 
outlines).  You also might want to specifically ask regarding the 
geometry issues i pointed out earlier.

You can be pretty sure the original producer of this data set has this 
information and if there is interest in having this data in OSM they 
should also be willing to provide such information.

In OpenStreetMap we put high importance on knowing and documenting how 
data is acquired since - as every experienced mapper knows - sources of 
information can be faulty and misleading.  Just because someone says: I 
have this data here and you may use it and it mostly looks reasonable 
and plausible at the first glance does not mean we should throw all our 
sense for critical evaluation of sources out of the window, especially 
if you plan to add several hundred thousand new features.

> The public 2011 data set is cut into 2×1 km tiles, slicing through
> buildings on the border. There are no common attributes which would
> allow repair, as the data is packaged in AutoCAD DWG files.

This alone should not be a problem since you know the location of the 
cuts and therefore could dissolve them based on position.

The real question is - is there a factual basis for the assumption that 
the data you intend to import is newer than the data in these files?  
Note a different level of detail is not necessarily an indication for 
age or accuracy of the data.

Christoph Hormann

More information about the Imports mailing list