[Imports] How good can an import be?

[Steve Singer]

On Tue, 5 Apr 2011, Richard Weait wrote:

> On Tue, Apr 5, 2011 at 7:55 AM, Mike N <niceman at att.net> wrote:
>> On 4/5/2011 4:58 AM, Andy Allan wrote:
>
>
> But the North American community has not grown like the European
> community did.  Is this because North Americans find OSM less
> appealing?  You suggest one possible reason this might be true.
>
>> The biggest issue in the US is that people's reaction is that we already 
>> have "Free Google, Bing, Yahoo and Mapquest" maps.  Why is there even 
>> another map?
>
> Possibly.  That is sometimes the reaction when I introduce people to OSM.

It isn't just the availability of free-beer map tiles like Google and Bing, 
North America has had free data available on a national scale for a while.


>
> By comparison, consider Toronto, pre-imports.  See the Geofabrik
> animation, around August 2007.
>
> http://www.geofabrik.de/gallery/history/index.html#toronto
>
> It seems to me, from memory, that starting around August 2007 was a
> boom-time of new OSM contributors in Toronto.  We had some of the city
> grid on the map, and a few detailed neighbourhoods.  A new user faced
> with that might say, "I'll come back when your done" but a new mapper
> might just say, "Oh, look, I can add my neighbourhood and make it look
> better, like $nearby-neighbourhood."  That obviously-incomplete map
> may well have inspired dedicated mappers to start to participate.
> Imports didn't arrive in Canada until later in the animation.
>
> This is not a scientific study, obviously.
>

Toronto is an interesting example in that by the time GeoBase Road import 
made it to Toronto (April/May of 2009) most of the roads were already on the 
map (Toronto proper, north of the city in Vaughan northwards this wasn't the 
case).  I don't think new mappers are joining in the Toronto area at the 
rate they were in 2007 but I don't think you can attribute that to an 
import.

You also have to consider quality.  A lot of the roads that show up on that 
animation as being added around the 2008/early 2009 timeframe were what you 
call 'zero star edits'[1].  People (often from a different continent) were 
using Yahoo imagery to trace in an outline of a road that would be unamed and often 
be a single way encompassing multiple real winding roads.  That practice 
like poorly done imports produces a map that 'looks complete' at low zoom 
levels but isn't all that great. Some of the mapping yourself and others did 
was great but the low zoom map at the time, like a TIGER only low-zoom map 
can be a bit deceiving.



> As another, single point of comparison of communities, consider local
> OSM groups.
> http://usergroups.openstreetmap.de/
>
> There appears to be a local OSM user group for each
> 100 Million USA-ians.
> And one for each 10 Million Canadians.
> And one for each 2 Million Germans.
>
> The USA got imports early.
> Canada got imports a bit later.
> Imports seem much less popular in Germany.
>
> Correlation? Causation?  I don't know.

How does population density figure into this? In order for a user group to 
be viable you need a critical mass of mappers to be close enough together. 
The population in the USA is much more distributed across the country than 
in Canada where of the people live close to the US border, and the 
OSM user groups in canada tend to be focused.


> There was no objection to TIGER import when Dave did it.  As I recall,
> mappers asked him to import their counties even if it meant removing
> their work first.  We did not suspect then, what we suspect now.
>

Often in these discussions about imports we talk about 'mappers' but we 
don't talk about the users of the map data.  Collecting map data is fun but 
most of us do it with the expectation that someone is going to actually use 
that data.  Back in 2008 when I loaded the OSM data around Toronto onto my 
smartphone it was pretty useless as a map to navigate with (if my plan for 
the day was to get from a to b, vs collect data).  Even last 
summer I had loaded OSM data onto my GPS for a trip to Quebec (no import at 
the time) and I had to resort to a paper gas-station map because the roads 
just weren't on the map (and many that were had no road names).



> Now by comparison, imagine that your town has a mapper in every
> neighbourhood. How long would a new store on your main street go
> unmapped?  A new subdivision in the next neighbourhood?  I think that
> model updates faster and better than waiting for somebody else to
> collect and publish the data, then an import / merge / update to OSM.
>
> I think that local mapper, keeping their area of interest up to date,
> is the great strength of OSM.  We see it in Germany.  It's a much more
> localized strength in North America.
>
> And I think imports have limited that strength in North America.

Have other countries *outside* of Western Europe, especially with population 
densities comparable to North America that have had no imports been able to 
develop better communities than in North America?  Yes the English and 
Germans are awesome but other countries that have avoided imports might make 
for a more realistic comparison.


[1] - http://weait.com/content/how-well-can-you-map

Steve