---
title: Open Massachusetts Transportation Data
teaser:
tags: news
author: Dan Croak
published_on: 2009-12-12
---
thoughtbot is hosting the [Great American
Hackathon](http://www.sunlightlabs.com/hackathon09) at our Boston office today
and tomorrow. The focus of our group is on transportation data from the
[Massachusetts Bay Transportation Authority](http://mbta.com) (MBTA).
Like many other government agencies, the public loves to trash the MBTA for its
actual and perceived failures while simultaneously taking advantage of its
services at levels of about [400 million riders a
year](http://www.mbta.com/about_the_mbta/news_events/?id=15517&month=&year=).
In response, the agency's knee-jerk reaction has been to historically be less
transparent.
I believe it is important as application developers to support [an effort
underway](http://www.eot.state.ma.us/developers/) by the agency to reverse that
trend. We have a window of about a year (until November, 2010) to show
government officials what is possible when they provide open data to the
public.
Here's my current understanding of the lay of the land and some ideas of things
to hack on.
## Who's who
MassDOT was created November, 2009. It combines almost all transportation
departments in Massachusetts, such as:
* Registry of Motor Vehicles
* Highways
* FastLANE
The T (MBTA) still has a complicated legal relationship with MassDOT. They are
kind of semi-private. I imagine them as similar to the Federal Reserve or [The
Pentavirate](http://www.youtube.com/watch?v=TPMS6tGOACo).
The main real-time data provider is a vendor the government has hired,
[NextBus](http://www.nextbus.com).
## Outreach
The main resource for developers is the [MassDOT Developers
Page](http://www.eot.state.ma.us/developers). There's also a [Twitter
account](http://twitter.com/massdotdev), regular monthly meetings with
developers, and occasional contests:
This was the winning
[visualization](http://www.massdotdevelopersconference09.com/visualization)
from a contest last month.
## The pilot program
For our purposes, the main thing we care about is real-time XML feed they are providing as part of
a one year pilot program. It is possible that feed will be turned off November,
2010 if developers don't create anything interesting or Those Fatcats don't see
value in transparency.
Legally, they are required to follow a particular procurement system.
Cannot change length of the procurement process but developers and the press
can push for more lines to open up.
## What data is available
There's static data and real-time data.
## Static data
The static data is available on the [MassDOT Developers
page](http://www.eot.state.ma.us/developers/). The current zip file is from
December 2:
The data is in Google Transit Feed Spec (GTFS) format, for interesting reasons.
The Federal Transit Administration been holding meetings about standard data
formats for a long time, then Google came along and just did it. Now pretty
much every agency (the MBTA in Boston, the MTA in New York, etc.) can export
into GTFS.
Some of the other data such as station locations and routes is in [Keyhole
Markup Language](http://en.wikipedia.org/wiki/Keyhole_Markup_Language) (KML)
files.
Perhaps the best way to explore this data is through
[danchoi/openmbta](http://github.com/danchoi/openmbta), an open source Rails
app and iPhone app. Some of things it contains are:
* stops for vehicles ("Alewife Station", "Broadway Station", "Cleveland Circle",
etc.)
* data for subways, buses, ferries, commuter rails
The app is live at [OpenMBTA.org](http://openmbta.org/).
## real-time data
Real-time data is what everyone is interested in.
MassDOT Highways have a [daily XML feed for planned construction
events](http://www.eot.state.ma.us/developers/downloads/boscon.xml), answering
the question, "which roads are closed?".
The Registry of Motor Vehicles has a [branch wait-time XML
feed](http://www.eot.state.ma.us/developers/downloads/qmaticXML.aspx).
The MBTA has a [services advisory and updates RSS
feed](http://talerts.com/rssfeed/alertsrss.aspx).
The Granddaddy feed everyone is excited about, however, is the [MBTA real-time
XML feed](http://www.eot.state.ma.us/developers/realtime/).
> The trial feed includes data for bus route 39 which serves Jamaica
> Plain, the Longwood Medical Area, and Back Bay in Boston; and bus routes 111,
> 114, 116, and 117 which serve Haymarket Station, East Boston, Chelsea, and
> Revere.
## Example vehicle location request/response
Request:
http://webservices.nextbus.com/service/publicXMLFeed?command=vehicleLocations&a=mbta&r=39&t=1258125794429
Response:
## What apps are people building
**Wait time prediction**. [This
one](http://www.clickteam.info/davidn/mbta/mbtabuses.php) works but it's
interface is... not good. [This Mac OS X
widget](http://www.wirelust.com/mbta-bus-tracker-widget-for-osx/) looks better
but it's... a Mac OS X widget.
Currently, a major problem as I see it with this kind of app is that NextBus is
a vendor with a patented, proprietary system for calculating the wait time. If
the MBTA switches vendor or this genius system turns out to suck, customers of
a wait time app will be pissed.
I'd like to see this data supplemented with user-generated content (was the
train late or on time?) or see people write their own, open source algorithms
based on lat/long, stops, times, rush hour traffic, whatever.
**Vehicle location**. [This one](http://www.mbta-bus.appspot.com/) is nice &
simple. We haven't seen the ultimate interface for this, however. The cream of
the crop is currently the [CTA Bus
Tracker](http://www.ctabustracker.com/bustime/map/displaymap.jsp) in Chicago.
## What other apps could be built
**Tap out**. Unlike, say, the London Tube, the T is a "tap in" system. Riders
don't have to tap out to leave the system. Therefore, developers and trend
analysts are limited in their data sets of how people are using the T. A mobile
app could be built that allows riders to "tap out", simply there to built up
the data sets.
**Between stops API**. A very simple API could be layered on the real-time NextBus data that says
which stops the next vehicle is between. This could turn out to be a more
accurate prediction for certain riders, particularly commuters who are familiar
with the line.
**Demographics**. One thing that isn't clear for application developers is: who
is the customer?. For something as broad as public transportation, there is a
wide variety of users, such as:
* college students
* the eldery
* the handicapped
* daily commuters
* tourists
They all could have very different needs. A simple user-generated application
to build demographics associated with each line could be very useful for
developers of future applications. For example, I ride the Red Line and the 1
bus. I know when commuters to downtown Boston ride it, when MIT and Harvard
students ride it, etc.
## Where to go next
Sign up for the [MassDOTDevelopers Google
Group](http://groups.google.com/group/massdotdevelopers).
Talk to local news stations like WBZ, NECN, WBUR, Fox 25, and boston.com. They
already add value to the government data by flying helicopters, watching
traffic cameras, and passing that along to their customers. There's potentially
more data at those businesses that could be added to the mix. A company called
[Smart Route Systems](http://www.smartroute.com) is often the point company for
those news stations.
Join us today and tomorrow at our Boston office as we hack together an app or
two!