<<

MBTA-realtime Integrating predictions and alerts into one GTFS-based platform

T3 Open Data Webinar, December 2013 Dave Barker, [email protected] Manager of Operations Technology, MBTA Motivation Vision Execution Results Response Lessons The MBTA:

• Serves 175 member cities and towns • Carries 1.3 million passengers each weekday • Operates heavy rail, , , , , and

• Extensive data sharing: GTFS, realtime, more Image credit: walkingbostonian.blogspot.com 2009-2012: Rapid development led to numerous feeds

GTFS GTFS

API Bus (NextBus) GTFS- Bus (RealTimeBus) realtime csv/json/ Commuter rail xml

csv/json Heavy rail 2.0 csv/json/ Heavy rail 1.0 xml

Alerts 1.0 RSS

Motivation Vision Execution Results Response Lessons The MBTA-realtime vision Bus locations API NextBus (NextBus) Limited GTFS

GTFS GTFS

Com.rail predictions GTFS- New realtime Subway predictions MBTA-realtime software API (XML, JSON) status RSS Alerts thru GUI (alerts only) Phase I, June 2013 Bus locations API NextBus (NextBus) Limited GTFS

GTFS GTFS (schedule only)

Com.rail predictions GTFS- New realtime Subway predictions MBTA-realtime software API (XML, JSON) Elevator status RSS Alerts thru GUI (alerts only) Motivation Vision Execution Results Response Lessons Technical details

• Designed and written by IBI Group for MBTA • C# application with Microsoft SQL Server back-end on two Amazon cloud servers • Based around foundation of GTFS data • Alert GUI • XML / JSON API, GTFS-realtime, RSS • Website, subscription service Alert GUI

Today’s trains (per GTFS)

Alert will clear after selected train scheduled to reach destination

Text generated as you make selections Recurring alerts

Details and URL Developer output & outreach

Motivation Vision Execution Results Response Lessons MBTA to Customers (SMS, email, website) Developers to customers (web, apps, notification) Google Transit BOS Roadify Developers to customers (web, apps, notification) Embark ProximiT Motivation Vision Execution Results Response Lessons Customers

#1 Complaint #2 Complaint I want I want fewer more alerts! alerts! Customers Positive - 30,000 subscribers - Alert detail & precision - Web page design Negative - Consistency - SMS delays - Subscriber options Internal users

Positive - Simple alerts are easy - Having accurate lists of trips, stops - Ability to review user actions Negative - Complex alerts are hard - Many audiences to consider at once (Alerts vs. website, email vs. SMS…) - Some repetitive actions Developers Positive - 100+ have registered to develop (!) - 30 of 65 polled “Plan to release app” - Alerts integrated quickly into some of the most popular apps Negative - RSS feed “good enough” for many - Many holding off until phase II (real-time) - No “T-Alert service killer” app yet Developers’ Plans using MBTA Data (November 2013 Survey)

Using in released app Using in app in development Might use Won't use

7 9 13 13 13 12

19 8 20 23 9 7 29 9 12

17

23 11 20 16 13 20

16 16 6 12 14 2 5 10 9 8 5 6 5 5 2 3 0 0 Alerts: RSS Schedule: GTFS Schedule: Commuter connect Commuter MBTA-realtime API MBTA-realtime Alerts: GTFS-realtime Subway predictions: v2 predictions: Subway v1 predictions: Subway Commuter rail predictions rail Commuter Bus predictions: NextBus API NextBus Bus predictions: Bus predictions: GTFS-realtime predictions: Bus Motivation Vision Execution Results Response Lessons Lessons •GTFS base successful •Development method worked well •Developer encouragement •Real-time information more fun for developers than alerts •The advantage of third parties

Next steps •Real-time integration for subway, commuter rail •Twitter integration •Steps to reduce message volume and improve formatting •Encourage further development •Leverage the API ourselves Thank you Visit realtime.mbta.com for more.

Acronym Reference - API Application Programming Interface - CSV Comma-Separated Value - GTFS General Transit Feed Specification - GUI Graphical User Interface - JSON JavaScript Object Notification - MBTA Mass. Bay - RSS Really Simple Syndication - SMS Short Message Service - XML eXtensible Markup Language

Dave Barker, Manager of Operations Technology, [email protected]