FW: OpenTripPlanner?
Kevin Chambers
This is one in a series of emails I'm forwarding to the new GTFS Flexible Transit Working Group. These emails make up the initial conversation that led to the creation of the working group. I'm sending them with permission of the original senders for archival purposes, allowing new participants to the conversation to get up to speed.
From: Kevin Webb [mailto:kw...@conveyal.com]
Sent: Thursday, April 18, 2013 12:27 PM
To: Denis Haskin
Cc: Aaron Antrim; Sean Barbeau; Heather Menninger; Becker, Jeff; Kevin Chambers; Roger Teal; David Emory
Subject: Re: OpenTripPlanner?
Denis,
I didn't talk about this in the context of yesterday's call, however, we (Conveyal) are curious about OTP applicability in more "informal" transit contexts. We have a pilot under way in Mexico City this spring where we'll be adapting OTP use a modified GTFS
feed for systems that aren't stop-based (this includes much of the world's transit). This isn't quite the DRT use case we talked about yesterday but it's a step in that direction. We'll be inviting folks to join a similar GTFS-"informal" working group this
spring as we hash out the data standards. There may be areas where this overlaps with the DRT conversation, but for the most part it is just a simplification of the existing GTFS spec to remove the burden of creating stops/schedules that don't actually exist
in more informal systems.
Also, perhaps more relevant, there's a startup in NYC that's experimenting with/contributing to OTP around dynamic dispatch applications for a different user community. I'm not sure how far along they are or if they're willing to collaborate but as far as I know they're actively working on this kind of problem. I've been meaning to reach out and check on their progress. I'll do that and let you know if they're interested in joining the dialog.
The big question in the DRT case is how well you can model vehicle routing. As a multimodal planner OTP supports this, however, it's largely a data problem. Road-based routing is less complex than transit routing but is only successful to the degree that your input road network is correct. We use OSM for everything OTP does on street, so unless you're planning to leverage a different data source any DRT/OTP application would require evaluating OSM in the target region. That said, the less you care about precise turn-by-turn directions the more likely that OSM will be sufficient.
Hope this helps!
Kevin
On Thu, Apr 18, 2013 at 3:10 PM, Denis Haskin <DHa...@camsys.com> wrote:
On the call yesterday several people expressed interest in using OpenTripPlanner with non-fixed-route trip planning. Who was it? Is anyone actually doing any work yet in this area? Jeff, I know you said you folks are doing it to some degree with your own internal trip planner.
Thanks,
--
Denis Haskin
Senior Architect/Senior Associate
Cambridge Systematics, Inc
direct line: 617-234-0480
www.camsys.com
--
Kevin Chambers
This is one in a series of emails I'm forwarding to the new GTFS Flexible Transit Working Group. These emails make up the initial conversation that led to the creation of the working group. I'm sending them with permission of the original senders for archival purposes, allowing new participants to the conversation to get up to speed.
From: Barbeau, Sean [mailto:bar...@cutr.usf.edu]
Sent: Thursday, April 18, 2013 12:33 PM
To: Kevin Webb; Denis Haskin
Cc: Aaron Antrim; Heather Menninger; Becker, Jeff; Kevin Chambers; Roger Teal; David Emory
Subject: RE: OpenTripPlanner?
Denis,
We haven’t done much development work ourselves on OpenTripPlanner server-side (we focus more on mobile), but we did set up an OTP instance in Tampa as part of a research project a year or two ago that includes 3 regional agencies (HART – Tampa, PSTA – Pinellas County, USF RullRunner campus shuttle):
http://opentripplanner.usf.edu/
While we wouldn’t have the resources to do the OTP software development to support non-fixed-route transit ourselves, we could probably help with testing OTP modifications made by others to support DRT service for agencies like HART and Lynx, whose data we could scrape together to test a new GTFS+DRT format.
Sean
Kevin Chambers
This is one in a series of emails I'm forwarding to the new GTFS Flexible Transit Working Group. These emails make up the initial conversation that led to the creation of the working group. I'm sending them with permission of the original senders for archival purposes, allowing new participants to the conversation to get up to speed.
From: Denis Haskin [mailto:DHa...@camsys.com]
Sent: Thursday, April 18, 2013 1:11 PM
To: Kevin Webb
Cc: Aaron Antrim; Sean Barbeau; David Emory; Heather Menninger; Becker, Jeff; Kevin Chambers; Roger Teal
Subject: Re: OpenTripPlanner?
Lots of good information -- thanks, Kevin and Sean.
We definitely don't care about detailed directions within the flex part of the routing. We're expecting to do it a much coarser level. For example, a veteran in some Atlanta suburb who needs to get to the VA; what are his/her options, very broadly, so they can evaluate them by time, cost, ease, etc: fully fixed route vs. maybe volunteer ride plus fixed route, vs. dial-a-ride, etc. It's possible this is completely beyond the scope of OTP, but I'm hoping not.
And as I think I mentioned on the call, we're not handling booking at all right now... so we would be telling someone their options, but for some of the options there would be lots of caveats--e.g. that there may not be capacity, you have to book.
The GTFS-"informal" work sounds definitely related to what we all are starting here with DRT, so I hope (and am sure) there will be collaboration. If that NYC startup is willing to reveal some of what they're doing, that would be great.
And completely out of the scope of this discussion, but just fyi one of the big things for the ARC VTCLI project (and this has also come up with other RFPs) is the requirement to capture when a trip *can't* be made. That is, documenting the lack of services and capacity is incredibly important... perhaps almost of equal importance with providing trip planning itself.
--
Denis Haskin
Senior Architect/Senior Associate
Cambridge Systematics, Inc
direct line: 617-234-0480
www.camsys.com
-----Kevin Webb <kw...@conveyal.com> wrote: -----
To: Denis Haskin <DHa...@camsys.com>
From: Kevin Webb <kw...@conveyal.com>
Date: 04/18/2013 03:28PM
Cc: Aaron Antrim <aa...@trilliumtransit.com>, Sean Barbeau <bar...@cutr.usf.edu>, Heather Menninger <hea...@ammatransitplanning.com>,
"Becker, Jeff" <Jeff....@rtd-denver.com>, Kevin Chambers <kcha...@rideconnection.org>, Roger Teal <Roger...@demandtrans.com>,
David Emory <dem...@conveyal.com>
Subject: Re: OpenTripPlanner?
Kevin Chambers
This is one in a series of emails I'm forwarding to the new GTFS Flexible Transit Working Group. These emails make up the initial conversation that led to the creation of the working group. I'm sending them with permission of the original senders for archival purposes, allowing new participants to the conversation to get up to speed.
From: Denis Haskin [mailto:DHa...@camsys.com]
Sent: Thursday, April 18, 2013 1:19 PM
To: Barbeau, Sean
Cc: Aaron Antrim; David Emory; Heather Menninger; Becker, Jeff; Kevin Chambers; Kevin Webb; Roger Teal
Subject: RE: OpenTripPlanner?
Thanks --
--
Denis Haskin
Senior Architect/Senior Associate
Cambridge Systematics, Inc
direct line: 617-234-0480
www.camsys.com
-----"Barbeau, Sean" <bar...@cutr.usf.edu> wrote: -----
To: Denis Haskin <DHa...@camsys.com>
From: "Barbeau, Sean" <bar...@cutr.usf.edu>
Date: 04/18/2013 04:17PM
Cc: Aaron Antrim <aa...@trilliumtransit.com>, David Emory <dem...@conveyal.com>, Heather Menninger <hea...@ammatransitplanning.com>,
"Becker, Jeff" <Jeff....@rtd-denver.com>, Kevin Chambers <kcha...@rideconnection.org>, Kevin Webb <kw...@conveyal.com>,
Roger Teal <Roger...@demandtrans.com>
Subject: RE: OpenTripPlanner?
Yes, all current service from HART, PSTA, and Bull Runner in OTP is fixed route.
We know that HART runs Flex-Route service, but it’s not currently reflected in their GTFS data:
HART and PSTA (via contractor) also provide paratransit service.
Sean
From: Denis Haskin [mailto:DHa...@camsys.com]
Sent: Thursday, April 18, 2013 4:12 PM
To: Barbeau, Sean
Cc: Aaron Antrim; David Emory; Heather Menninger; Becker, Jeff; Kevin Chambers; Kevin Webb; Roger Teal
Subject: RE: OpenTripPlanner?
Sean -- those are all fixed-route, yes?
And good to hear about the potential to provide some GTFS/DRT data from other agencies, when we get there.
Thanks,
--
Denis Haskin
Senior Architect/Senior Associate
Cambridge Systematics, Inc
direct line: 617-234-0480
www.camsys.com
-----"Barbeau, Sean" < bar...@cutr.usf.edu> wrote: -----
To: Kevin Webb <kw...@conveyal.com >, Denis Haskin <DHa...@camsys.com
>
From: "Barbeau, Sean" < bar...@cutr.usf.edu>
Date: 04/18/2013 03:33PM
Cc: Aaron Antrim < aa...@trilliumtransit.com>, Heather Menninger < hea...@ammatransitplanning.com>, "Becker, Jeff" <
Jeff....@rtd-denver.com>, Kevin Chambers < kcha...@rideconnection.org>, Roger Teal < Roger...@demandtrans.com>, David Emory <
dem...@conveyal.com>
Subject: RE: OpenTripPlanner?
Denis,
We haven’t done much development work ourselves on OpenTripPlanner server-side (we focus more on mobile), but we did set up an OTP instance in Tampa as part of a research project a year or two ago that includes 3 regional agencies (HART – Tampa, PSTA – Pinellas County, USF RullRunner campus shuttle):
http://opentripplanner.usf.edu/
While we wouldn’t have the resources to do the OTP software development to support non-fixed-route transit ourselves, we could probably help with testing OTP modifications made by others to support DRT service for agencies like HART and Lynx, whose data we could scrape together to test a new GTFS+DRT format.
Sean
From: Kevin Webb [mailto:kw...@conveyal.com
]
Sent: Thursday, April 18, 2013 3:27 PM
To: Denis Haskin
Kevin Chambers
This is one in a series of emails I'm forwarding to the new GTFS Flexible Transit Working Group. These emails make up the initial conversation that led to the creation of the working group. I'm sending them with permission of the original senders for archival purposes, allowing new participants to the conversation to get up to speed.
From: Heather Menninger [mailto:hea...@ammatransitplanning.com]
Sent: Thursday, April 18, 2013 2:26 PM
To: 'Denis Haskin'; 'Kevin Webb'
Cc: 'Aaron Antrim'; 'Sean Barbeau'; 'David Emory'; 'Becker, Jeff'; Kevin Chambers; 'Roger Teal'
Subject: RE: OpenTripPlanner?
Denis this issues and items, including identifying when a trip “can’t” be made are of great concern here to in this Inland Empire southern California region. We are trying to identify those and develop – or pose certain volunteer mileage reimbursement – work arounds. So I heartily concur with what you raise and hope for in terms what are his/her options, recognizing that it may or may not be possible.
Is this what you all meant by “trip discovery”?
-Heather
Heather Menninger, Principal
[Heather Menninger Visscher]
AMMA Transit Planning
393 Two Trees Road, Riverside, CA. 92507
v. 951 784-1333 f. 951 784-1212
From: Denis Haskin [mailto:DHa...@camsys.com]
Sent: Thursday, April 18, 2013 1:11 PM
To: Kevin Webb
Cc: Aaron Antrim; Sean Barbeau; David Emory; Heather Menninger; Becker, Jeff; Kevin Chambers; Roger Teal
Subject: Re: OpenTripPlanner?
Lots of good information -- thanks, Kevin and Sean.
We definitely don't care about detailed directions within the flex part of the routing. We're expecting to do it a much coarser level. For example, a veteran in some Atlanta suburb who needs to get to the VA; what are his/her options, very broadly, so they can evaluate them by time, cost, ease, etc: fully fixed route vs. maybe volunteer ride plus fixed route, vs. dial-a-ride, etc. It's possible this is completely beyond the scope of OTP, but I'm hoping not.
And as I think I mentioned on the call, we're not handling booking at all right now... so we would be telling someone their options, but for some of the options there would be lots of caveats--e.g. that there may not be capacity, you have to book.
The GTFS-"informal" work sounds definitely related to what we all are starting here with DRT, so I hope (and am sure) there will be collaboration. If that NYC startup is willing to reveal some of what they're doing, that would be great.
And completely out of the scope of this discussion, but just fyi one of the big things for the ARC VTCLI project (and this has also come up with other RFPs) is the requirement
to capture when a trip *can't* be made. That is, documenting the lack of services and capacity is incredibly important... perhaps almost of equal importance with providing trip planning itself.
--
Denis Haskin
Senior Architect/Senior Associate
Cambridge Systematics, Inc
direct line: 617-234-0480
www.camsys.com
From: Kevin Webb <kw...@conveyal.com>
Date: 04/18/2013 03:28PM
Cc: Aaron Antrim <aa...@trilliumtransit.com>, Sean Barbeau <bar...@cutr.usf.edu>, Heather Menninger <hea...@ammatransitplanning.com>,
"Becker, Jeff" <Jeff....@rtd-denver.com>, Kevin Chambers <kcha...@rideconnection.org>, Roger Teal <Roger...@demandtrans.com>,
David Emory <dem...@conveyal.com>
Subject: Re: OpenTripPlanner?
Denis,
I didn't talk about this in the context of yesterday's call, however, we (Conveyal) are curious about OTP applicability in more "informal" transit contexts. We have a pilot under way in Mexico City this spring where we'll be adapting OTP use a modified GTFS
feed for systems that aren't stop-based (this includes much of the world's transit). This isn't quite the DRT use case we talked about yesterday but it's a step in that direction. We'll be inviting folks to join a similar GTFS-"informal" working group this
spring as we hash out the data standards. There may be areas where this overlaps with the DRT conversation, but for the most part it is just a simplification of the existing GTFS spec to remove the burden of creating stops/schedules that don't actually exist
in more informal systems.
Also, perhaps more relevant, there's a startup in NYC that's experimenting with/contributing to OTP around dynamic dispatch applications for a different user community. I'm not sure how far along they are or if they're willing to collaborate but as far as I know they're actively working on this kind of problem. I've been meaning to reach out and check on their progress. I'll do that and let you know if they're interested in joining the dialog.
The big question in the DRT case is how well you can model vehicle routing. As a multimodal planner OTP supports this, however, it's largely a data problem. Road-based routing is less complex than transit routing but is only successful to the degree that your input road network is correct. We use OSM for everything OTP does on street, so unless you're planning to leverage a different data source any DRT/OTP application would require evaluating OSM in the target region. That said, the less you care about precise turn-by-turn directions the more likely that OSM will be sufficient.
Hope this helps!
Kevin
On Thu, Apr 18, 2013 at 3:10 PM, Denis Haskin < DHa...@camsys.com> wrote:
On the call yesterday several people expressed interest in using OpenTripPlanner with non-fixed-route trip planning. Who was it? Is anyone actually doing any work yet in this area? Jeff, I know you said you folks are doing it to some degree with your own internal trip planner.
Thanks,
--
Denis Haskin
Senior Architect/Senior Associate
Cambridge Systematics, Inc
direct line: 617-234-0480
www.camsys.com
--
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2013.0.3272 / Virus Database: 3162/6253 - Release Date: 04/18/13
Kevin Chambers
This is one in a series of emails I'm forwarding to the new GTFS Flexible Transit Working Group. These emails make up the initial conversation that led to the creation of the working group. I'm sending them with permission of the original senders for archival purposes, allowing new participants to the conversation to get up to speed.
From: Roger Teal [mailto:Roger...@demandtrans.com]
Sent: Thursday, April 18, 2013 3:50 PM
To: Kevin Webb; Denis Haskin
Cc: Aaron Antrim; Sean Barbeau; Heather Menninger; Becker, Jeff; Kevin Chambers; David Emory
Subject: RE: OpenTripPlanner?
Kevin:
This is great stuff. I have a couple of colleagues who have done lots of consulting for World Bank in locations throughout the world, and whose presentations have educated me as to the nature of “fixed” route systems in Latin America, Africa, and Asia, and it would be terrific if the GTFS spec could be adapted to handle the data situations that those services present. As you know, mobile devices are hugely important as the web-based interface for most folks in those emerging market economies, and if an OTP-based mobile app for those environments can be piggybacked on top of a good data representation of the services as they actually exist and operate, I would think it would be highly beneficial. Include me on any communication about the GTFS informal working group on this subject and I will make sure those folks are brought into the loop, they could be great contributors. They are more planners than technology folks, but have an appreciation of how technology can be a game changer.
I assume that OSM stands for OpenStreetMaps. Is there an API that is commonly used for accessing OSM data? I see that the OSM API is just for editing the underlying map data, not for retrieving it for use in a map-based app (or so is my understanding).
Roger
From: Kevin Webb [mailto:kw...@conveyal.com]
Sent: Thursday, April 18, 2013 2:27 PM
To: Denis Haskin
Kevin Chambers
This is one in a series of emails I'm forwarding to the new GTFS Flexible Transit Working Group. These emails make up the initial conversation that led to the creation of the working group. I'm sending them with permission of the original senders for archival purposes, allowing new participants to the conversation to get up to speed.
From: Barbeau, Sean [mailto:bar...@cutr.usf.edu]
Sent: Friday, April 19, 2013 6:21 AM
To: Roger Teal; Kevin Webb; Denis Haskin
Cc: Aaron Antrim; Heather Menninger; Becker, Jeff; Kevin Chambers; David Emory
Subject: RE: OpenTripPlanner?
Roger,
Yes, OpenStreetMap does support programmatic access to data via an API. Info on that OSM API is here:
http://wiki.openstreetmap.org/wiki/Xapi#Servers
In the default OpenTripPlanner setup, OTP grabs a region from OSM based on the area of coverage by the GTFS data. Actual code that does that is here:
The OSM API servers can be slow depending on demand, so there are some sites that do an extract and mirror of particular OSM areas that are downloadable as files. One such site is CloudMade:
http://downloads.cloudmade.com/
For smaller areas, you can actually do an export via the Openstreetmap.org user interface (look at menu header on that page).
More information about how OTP builds a routable graph from OSM and other sources (including NED data) is here:
https://github.com/openplans/OpenTripPlanner/wiki/GraphBuilder
Sean