Hey Raines,
Thanks for bringing this up. The question of crowdsourcing transit
data is an interesting one--a lot of people have talked about it over
the years, but I don't know of many efforts that have succeeded:
* The original late-90's transitinfo.org and SocalTIP websites, where
volunteers typed up transit schedules and put them up in text form.
* The old Métro program for PDAs, for which the authors solicited
user-contributed data files for subway systems. These files were very
simple (most didn't even have station locations!), and the authors
claimed copyright over the submitted data.
* Most transit mashups are usually either a single person either
hand-entering a small amount of subway information, or scraping a
larger amount of bus information from the agency's site, so it's
usually a very small crowd. OnNYTurf has done some interesting stuff
with having users send in photos of station maps, and also with their
wiki for station metadata (though even then it appears that 2 people
do most of the data entry there).
* There was an interesting effort that Kieran Huggins and Kevin
Branagan were working on following the original Toronto
TransitCamp--they scraped the TTC website and then had a mashup where
contributors could improve the extracted shapes of the routes to
better match the real schedules. I don't know if that software is
still under development, though.
Would Google use crowdsourced transit data for routing? It's not out
of the question, particularly if they thought that the group that put
it together was in a position to keep it accurate and current over the
years. But in most cases, I suspect the effort of crowdsourcing data
would be better spent on just convincing the agencies to share the
data that's already in their databases, or on building other useful
transit tools.
Joe
After thinking about it a little more, there is one part of transit
data that would be well-suited to crowdsourcing: collecting the shapes
for routes. Many agencies that are already sharing data with Google
and the public still aren't providing information on the paths that
the buses and trains take, either because they don't have the
information, or because their shape data has IP constraints. These
systems have to be shown by just drawing straight lines between the
stop locations, which can look bad in many cases.
So having good shapes to add to the schedule data would really help.
Not only that, but I think that creating route shapes on a map is
fairly rewarding for an individual contributor, so it's a good
candidate for an addictive community project.
Kieran just posted an update about the OpenTransit software I mentioned:
http://groups.google.com/group/transit-developers/browse_thread/thread/ea5354b102465741
Perhaps this tool, or the similar one that Chris St. Pierre has been
building for his Cincy Metro Interactive Map project, would make a
good starting point for such an effort.
Joe
Rather than straight lines, perhaps the line could conform to the road
line.
Some people say Wikipedia is more accurate than Encyclopedia Britannica. I am a big believer in “crowdsourcing” (glad we now have a name for it). I don’t see how it could hurt to create a site and let the user beware. We might be surprised how accurate it might turn out.