2013 UTC Southeast - OneBusAway – Sharing real time transit information via open-source software
OneBusAway Multi-region – Rapidly expanding mobile transit ......Apr 17, 2014 · Transit Data in...
Transcript of OneBusAway Multi-region – Rapidly expanding mobile transit ......Apr 17, 2014 · Transit Data in...
4/17/2014
1
Center for Urban Transportation Research | University of South Florida
OneBusAway Multi-regionRapidly Expanding Mobile Transit Apps to New Cities
Sean J. Barbeau, Ph.D.
2
Background
• Real-time transit information has many benefits– Shorter perceived wait time [1]
– Shorter actual wait time [1]
– Lowers learning curve for new riders [2]
– Increased ridership (maybe?) [3]
– Increased feeling of safety (e.g., at night) [5][6]
• Riders prefer accessing real-time transit info via mobile apps [1]
[1] Kari Edison Watkins, Brian Ferris, Alan Borning, G. Scott Rutherford, and David Layton (2011), "Where Is My Bus? Impact of mobile real-time information on the perceived and actual wait time of transit riders," Transportation Research Part A: Policy and Practice, Vol. 45 pp.839-848.[2] C. Cluett, S. Bregman, and J. Richman (2003). "Customer Preferences for Transit ATIS," Federal Transit Administration. Available at http://ntl.bts.gov/lib/jpodocs/repts_te/13935/13935.pdf#sthash.jwn5Oltr.dpuf[3] Lei Tang and Piyushimita Thakuriah (2012), "Ridership effects of real-time bus information system: A case study in the City of Chicago," Transportation Research Part C: Emerging Technologies, Vol. 22 pp. 146-161.[4] Aaron Steinfeld and John Zimmerman, "Interviews with transit riders in San Francisco and Seattle," ed, 2010.[5] Brian Ferris, Kari Watkins, and Alan Borning (2010), "OneBusAway: results from providing real-time arrival information for public transit," in Proceedings of the 28th International CHI Conference on Human Factors in Computing Systems, Atlanta, Georgia, USA, pp. 1807-181[6] A. Gooze, K. Watkins, and A. Borning (2013), "Benefits of Real-Time Information and the Impacts of Data Accuracy on the Rider Experience," in Transportation Research Board 92nd Annual Meeting, Washington, D.C., January 13, 2013.
Tony Kurdzuk/The Star-Ledger
4/17/2014
2
3
Challenge
• Custom mobile apps are expensive– Some estimates > $150,000k for a single platform [7]
– Doesn’t include maintenance cost
• Implementing and maintaining custom apps on all popular smartphone platforms is cost-prohibitive
• Open data is good, but doesn’t always result in apps by 3rd party developers
• How can we cost-effectively launch mobile transit apps in new cities?
[7] Larry Lauvray "iPad App Development Cost - A Breakdown." Propelics Blog, Vol. Accessed at: http://www.propelics.com/ipad-app-development-cost-a-breakdown/
4
OneBusAway
• Open-source real-time transit information system
• Originally developed at University of Washington– Ph.D. work of Brian Ferris and Kari Watkins– Used in multiple research studies to
understand how real-time info affects transit riders
• Includes server-side software– Process bulk real-time transit info– Provides Application Programming Interface
(API) for apps• Includes mobile apps for:
– iPhone– Android– Windows Phone– Windows 8
http://onebusaway.org
4/17/2014
3
5
OneBusAway – Mobile Apps
Android Windows Phone &Windows 8
iPhone
Support user location, route, stop contextual/personalized informationAll OPEN-SOURCE!
6
OneBusAway
• 2012 – Mobile apps only available in Puget Sound
• How do we expand them to new cities?
http://onebusaway.org
4/17/2014
4
7
Original OneBusAway mobile app design
8
Why don’t we duplicate the mobile apps for each city?
• Creates app maintenance work for each city, per platform
• Clutters app stores
• Fragments source code
4/17/2014
5
9
OneBusAway “multi-region” design
• Each region maintains its own OneBusAway server software, but uses the same mobile apps– Reduces burden on regions to maintain apps– Eliminates burden on a centralized entity to
maintain data quality control, servers, and customer support for each new region
– Doesn’t clutter app stores – all users download the same “OneBusAway” app for all regions
• “You bring the server, we bring the apps”
10
Multi-region architecture
• Region information is stored in centralized server directory (i.e., Regions API)
• Apps now find nearby regions using Regions API
4/17/2014
6
11
OneBusAway Server Directory
12
OBA – Android app
4/17/2014
7
13
New OneBusAway Region Checklist
Transit Data in GTFS format AVL system that provides arrival estimates Implement a GTFS-realtime (or SIRI) feed Set up a OneBusAway Server Do some quality-control testing Launch OneBusAway apps in new city!
• Via request to onebusway-developers group!
14
Adding OneBusAway regions to apps
• Following the request to the onebusaway_developergroup, the new region can be added
• Two stages for a new region:1. Experimental:
• Regions in development• Users can enable in iPhone and Android apps via “Settings->Enable
Experimental Regions”2. Production:
• Meets a certain standard in terms of quality and quantity of real-time data
• Region is available in all apps by default• See http://goo.gl/ydbc0c for details
4/17/2014
8
15
ONEBUSAWAY TAMPASetting up a new OneBusAway region – a case study
16
New OneBusAway Region Checklist
Transit Data in GTFS format AVL system that provides arrival estimates Implement a GTFS-realtime (or SIRI) feed Set up a OneBusAway Server Do some quality-control testing Launch OneBusAway apps in new city!
- Here’s where we started in Tampa…
4/17/2014
9
17
Creating a new region - Tampa
1. Establish a GTFS-realtime feed
GTFS-realtime software is open-source :https://github.com/CUTR-at-USF/HART-GTFS-realtimeGenerator/wiki
18
Creating a new region - Tampa
2. Configure OneBusAway software
GTFS-realtime software is open-source :https://github.com/CUTR-at-USF/HART-GTFS-realtimeGenerator/wiki
4/17/2014
10
19
GTFS-realtime format (in plain text)
trip_update { trip {
trip_id: "974372" } stop_time_update {
stop_sequence: 17 arrival {
delay: -180 uncertainty: 30
}stop_id: "571"
}vehicle {
id: "1301"}
}
Estimated Arrival Times (TripUpdate)
Data displayed in mobile apps
20
GTFS-realtime format (in plain text)
trip_update { trip {
trip_id: "974372" } stop_time_update {
stop_sequence: 17 arrival {
delay: -180 uncertainty: 30
}stop_id: "571"
}vehicle {
id: "1301"}
}
vehicle {trip {
trip_id: "974372"}position {
latitude: 28.02931 longitude: -82.456566bearing: 360.0speed: 0.0
}vehicle {
id: "1301"}
}
Estimated Arrival Times (TripUpdate) Vehicle Locations (VehiclePositions)
Data displayed in mobile apps
Not currently used by mobile apps
4/17/2014
11
21
OBA – Android app
Delay value
Arrival estimate, based on real-time delay
22
Challenges – Technology / Process
Challenge Solution / Lesson Learned
Old documentation for OneBusAway at the start of the project
Create new documentation!Painful at first, but helps others
Unknown accuracy of arrival times from vendor AVL
On-the-ground testing and validation necessaryAfter several iterations, got good results!
Communication with various parties• Agency ITS staff• Agency IT staff• Agency Marketing staff• Vendor
Get upper-management buy-in for Agency, have in-person meetings. Identify champion within agency.Takes time, but reduces overall turnaround time for requests
OneBusAway Tampa
4/17/2014
12
23
Challenge Solution / Lesson Learned
GTFS - arrival time for Stop A is after arrival time for Stop B during trip
Work with agency to resolve problemShow the agency error in software, create documentation to explain the current data issues with examples
GTFS - Seconds resolution (“HH:MM:SS” ) for time not included ”GTFS - Incorrect route URLs prevented app users from seeing schedules ”AVL - tripIDs not matching GTFS tripIDs Examine data dictionary for AVL system,
talk directly to vendorConference call helped resolve issue
AVL - Flipped +/- for delay value Extensive troubleshooting and isolation of issue, talk directly to vendorFinally resulted in on-site meeting, where we resolved the problem
AVL – Agency database replication issues ”AVL - Use “deviation” or “predicted_deviation” value?
On-the-ground testing and validation necessaryAfter several iterations, got good results!
Challenges – DataOneBusAway Tampa
24
Tampa - From pilot to production
• USF launched a pilot of OneBusAway in Tampa with approximately 200 users in February 2013.– Partnered with Georgia Tech to do controlled study of
impact of real-time info on riders– USF maintained OneBusAway for the pilot
• Public launch – August 2013– HART adopted OneBusAway Tampa, now runs on their
servers (tampa.onebusaway.org)
– USF assisted HART in preparing RFP for long-term vendor support (began February 2014)
4/17/2014
13
25
LESSONS LEARNED & CONCLUSIONS
The “take-away”
26
Lessons Learned
• Open-source software enabled this project – many contributors– Git and Github are huge enablers
• Volunteer developers are great for general advancement of the project, but for time-sensitive, or specific, tasks paid developers may be required
• Identifying an on-going funding source is important for limited centralized operations (e.g., review and merging of software contributions from community)
• Pattern of “tech demo -> production” region deployments seems to be emerging
4/17/2014
14
27
Conclusions
• OneBusAway mobile apps in new cities:– Tampa (tampa.onebusaway.org – Hillsborough Area Regional Transit)
– Atlanta (atlanta.onebusaway.org – Georgia Tech, w/ data from MARTA)
– Washington, D.C. (tech demo - http://oba.mobilitylab.org/ - Mobility Lab)
– York Region, Canada (tech demo – York Region Transit)
– (More soon)!
• OneBusAway multi-region -> easy to add new cities in the future– Sustainable, scalable, low-cost, and low-maintenance
model
28www.onebusaway.org
4/17/2014
15
29
Future work
• Ongoing research to understand impact of real-time info on riders– Helps us understand what users want, how to
encourage use of non-single occupancy vehicle modes– Justifies investments in the technology
• New features– Improved issue reporting– Contributions from deployments– OneBusAway for Google Glass– Others…
30
Acknowledgements
• Full acknowledgements for OneBusAway Multi-region Initiative are at:– http://bit.ly/OBA_multiregion_blog
• Funding for USF’s work provided by the National Center for Transit Research
• Hillsborough Area Regional Transit, for data and support of the OneBusAway Tampa public deployment
• University of Washington and Georgia Tech, for continued involvement in the project
• All OneBusAway community members