search

evz / magnificent-mileage

Open Trip Planner instance for Chicago
2 stars 3 forks source link

Bike directions suggest riding wrong way down one way street #7

Closed fgregg closed 11 years ago

fgregg commented 11 years ago

http://ec2-54-227-63-227.compute-1.amazonaws.com/opentripplanner-webapp/#/submit&fromPlace=41.921885,-87.718142&toPlace=41.888572,-87.636775&mode=BICYCLE&min=TRIANGLE&triangleTimeFactor=0&triangleSlopeFactor=0&triangleSafetyFactor=1&maxWalkDistance=4828&walkSpeed=1.341&time=8:48am&date=8/13/2013&arriveBy=false&itinID=1&wheelchair=false&preferredRoutes=&unpreferredRoutes=&bannedRoutes=&bannedStops=

On Lyndale

DJSekora commented 11 years ago

This is probably an issue with the OSM data. Although, it's conceivable that OTP assumes bikes can ignore oneways.

iandees commented 11 years ago

I checked the way in question and the data looks fine in OSM, but it was updated in June. @evz, where did you source the OSM data from? I wonder if the extract used doesn't have the most up-to-date data?

evz commented 11 years ago

Pretty sure it was from MapQuest. It's whatever OpenTripPlanner uses by default. Graph builder config is here but it's not obvious where it's actually downloading from. I'll dig deeper.

evz commented 11 years ago

Spoke too soon, it's here. And yes, it is MapQuest.

iandees commented 11 years ago

Yea, just found that file myself. MapQuest's jxapi instance is woefully out of date. Their version of the way I linked above they're at least a couple years old. I think it'd probably be better to use Mike's Chicago extract which would give us wider car routing coverage and is much newer.

iandees commented 11 years ago

Alternatively, change the base URL in OSMDownloader to http://jxapi.openstreetmap.org/xapi/api/0.6/. That instance is kept up to date and is probably faster, too.

evz commented 11 years ago

See #16