Results not showing the best options about modeify HOT 12 CLOSED

adding another trip for testing drive + transit bugs:

http://arlington.dev.conveyal.com/planner?from=2311%20N%20Powhatan%20St,%20Arlington,%20Virginia%2022205&to=1914%208th%20St%20NW,%20Washington,%20District%20of%20Columbia%2020001

This misses important trips unless car is disabled.

from modeify.

As of commit 6a708e491dfb74bc0252153ae3a599c7534c251a @trevorgerhardt 's original examples both return (red and (yellow or green)) as one of the top options.

from modeify.

@kpwebb your example also looks good in latest master (allowing strictly suboptimal trips). Activating or deactivating driving seems to return similar results, all of which look correct to me.

from modeify.

Looking better, but here's a strange one on drive to transit:

http://www.carfreeatoz.com/planner?from=1534%20Westbranch%20Dr,%20Mclean,%20Virginia%2022102&to=4922%20Brandywine%20St%20NW,%20Washington,%20District%20of%20Columbia%2020016

from modeify.

Ok, i'm seeing some additional examples that I asssume have to do with transit/street linkages:

From lorton VA we're doing great with bike to transit but fail plan the correct trip with drive to transit:

http://arlington.dev.conveyal.com/planner?from=Lorton%20Valley%20Rd,%20Lorton,%20Virginia,%20USA&to=1501%20Wilson%20Blvd,%20Arlington,%20Virginia,%20USA

Similar I'm not seeing correct correct drive to transit results for Silver Line trips in Tyson's Corner:

http://arlington.dev.conveyal.com/planner?from=8595%20Coral%20Gables%20Ln,%20Tysons%20Corner,%20Virginia%2022182&to=1501%20Wilson%20Blvd,%20Arlington,%20Virginia,%20USA

I think this is all about linkage given that we can clearly plan/pick the right trips for these locations.

from modeify.

Only some of the park and ride lots are linked into the graph at all because they are not mapped topologically (the lots just float over / under the roads). It is possible OTP just does not know where the lots are. Realistically people are not going to start mapping the lots with shared nodes so Laurent is fixing this right now.

from modeify.

The problem I mentioned above appears to already be fixed, just need to review and merge a pull request.

from modeify.

@kpwebb with the last discussion re: OSM park and rides at Tysons, are these trips still invalid or can this be closed for now?

from modeify.

@trevorgerhardt I need to check out the park and ride functionality anyway, so I'll report back here about whether it seems to be working.

from modeify.

http://localhost:5000/planner?from=44%20K%20St%20NW,%20Washington,%20District%20of%20Columbia,%20USA&to=33%20A%20St,%20Fort%20Mcnair,%20District%20of%20Columbia,%20USA&modes=BICYCLE,BUS,TRAINISH,WALK,CAR&start_time=7&end_time=9&days=M%E2%80%94F

This suggests three different ways to ride the wrong direction on the D4 and then reverse direction onto the X2/X9, while there are stops near the origin point on both the X routes. Weird.

from modeify.

More backward-travel. Is this because there's not a park and ride on the green/yellow line in this area, or it's not being correctly linked?

from modeify.

There are P&Rs at several Green/Yellow stations in that area and they seem to be properly linked in OSM. Here's a drive-to-transit trip that uses the College Park P&R (which appears to be closest to the start location in the above example):

http://arlington.dev.conveyal.com/planner?from=4811%20Ruatan%20St%2C%20College%20Park%2C%20Maryland%2020740&to=641%20S%20St%20NW%2C%20Washington%2C%20District%20of%20Columbia%2C%20USA&modes=TRAINISH%2CWALK%2CCAR&start_time=7&end_time=9&days=M%E2%80%94F

I was able to successfully plan drive-to-transit trips to the nearby Greenbelt and PG Plaza stations as well.

from modeify.