Martjin, You can get current information about how old/accurate the data are using the above link.
http://geogratis.gc.ca/api/en/nrcan-rncan/ess-sst/-/(urn:iso:series)geobase-national-road-network-nrn?sort-field=relevance To get the answer you need to look at the metadata: You select the area you are interested in (click) You need to view the full metadata (formatted NAP) (click) Look for the above tags under "dataQualityInfo" "DQ_AbsoluteExternalPositionalAccuracy" You will find 1:N Distance values that describe how accurate the data are. "EX_TemporalExtent" You will find 1:N timePosition values that describe how old the data are. The same should apply for the other layers (NHN, NRWN...), even if the distribution units may differ (whole country, watershed ...) Daniel -----Original Message----- From: Begin Daniel [mailto:[email protected]] Sent: Tuesday, 13 September, 2016 07:53 To: Martijn van Exel; Stewart C. Russell; [email protected] Subject: Re: [Talk-ca] Canvec attributes (roads) Bonjour Martijn AFAIK, here is a summary about how old and accurate is Canvec/GeoBase. Transport Layers? The roads are updated every 1-2 years for most of the provinces, 5-10 for others. Railways were updated 4-5 years ago over all the Canadian landmass. Other layers Water features are 5-40+ years old depending on the province/territory/latitude; Forest areas are 5-40+ years old depending on the province/territory/latitude [1]. The rest is 25-40+ years old. About the accuracy, the road network is about 10m (90%). The rest is usually better than 30m (90%) but you may find offsets between layers. Daniel [1] About forest areas, the latest GeoBase data results from images classification made about 5 years ago that gave the clumsy result Paul Ramsay recently shown on this list. The latest Canvec OSM tiles (2012?) had a mixture of old map data/new GeoBase data. -----Original Message----- From: Martijn van Exel [mailto:[email protected]] Sent: Monday, 12 September, 2016 23:06 To: Stewart C. Russell; [email protected] Subject: Re: [Talk-ca] Canvec attributes (roads) On 09/12/2016 06:50 PM, Stewart C. Russell wrote: > On 2016-09-12 04:08 PM, Martijn van Exel wrote: >> Aren't these files grouped by feature type? So if we look at roads we >> wouldn't also necessarily need to look at land use boundaries etc.? > > Canvec - the product supplied by NRCan to the general public - always > was split by feature type. It's the OSM tiles, of structure decided > long ago, that lump everything together. > > It's also available as effectively seamless FGDBs if you want to avoid > the cleanup required after using tiled data. The FGDBs retain the > critically important survey dates and accuracies - so you can easily > see how much data's 40 years old and has ±75 m positional accuracy. > Good to know. Are any of the transport related datasets that old or that inaccurate? I created an initial Canvec road network translation file for ogr2osm, so you can convert the Canvec shapefiles to OSM format easily (if you know how to work ogr2osm - let me know if you need help, but Paul Norman is the expert here!) It is located at https://github.com/mvexel/canvec-ogr2osm-translation/blob/master/canvec2016.py and I hope for many forks and improvements. Right now it does a basic job of translating the road classes to OSM types, and the most obvious attributes to the corresponding OSM tags. Let me know what you all think. Martijn _______________________________________________ Talk-ca mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-ca _______________________________________________ Talk-ca mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-ca _______________________________________________ Talk-ca mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-ca
