For public transit, travel time is the major constraint. The system will take you almost anywhere, assuming you have the time.
These maps visualize how long it will take to get there. Destinations are plotted from multiple origins based on an average estimated travel time. The maps retain compass directions from the origin, so a stop that is east of the origin is to the right, west of the origin is to the left. An estimated route for each stop also is shown.
The concept updates the familiar time map, using the MBTA’s online scheduling data. It can be refocused on any origin – in theory, even any origin contributed by a user (and can include user-contributed destinations too). The maps also could be used to show impacts from service changes, customized for the origin that is most of interest to the viewer. The original idea was to have a way to visualize the big impact of bridging the relatively small distance between the two halves of the rail system via the North-South Rail Link. The maps bring visual attention to the critical travel time and headway wait, for the entire system.
The maps are derived from a database, provided courtesy of MassDOT, containing official schedule times. The data was boiled down to two values: (1) how far apart are the stations or platforms, meaning how long a walk, and (2) how soon can you get from one station or platform to another using transit services. Walking speed was guesstimated at a leisurely two and a half miles per hour – as the crow flies, ignoring the street grid. Service times are averaged out and the number of trips between trip pairs are counted. Waiting times are the most complex calculation, and reflect an optimism that you can catch the service you want quickly, at least at the origin. Wait times were estimated at the lesser of half the headway or three minutes at the origin, increasing six minutes for every trip leg (up to nine minutes after the first trip/walk connection, fifteen minutes after two, etc). The idea is that riders will time their trips to avoid long headways, if they can. Half a minute is added for each connection.
The route lines are a bit quirky. Riders catch routes by “walking” to them from the origin or another stop. The walks cross the “white” space in the map. Some routes merge and converge at stops (and some stops on a line might be reached first by a different route). Those instances posed special problems. The maps address them by plotting the most frequent route or alternate route) on the local stops, with the other routes “hopping” to later stops that they serve. The skipping effect (and sometimes gapping effect) is most apparent on the Green Line and for some rail service that runs through rapid transit hubs. Most, and perhaps all of the gaps and skips in route plots reflect real-world conditions of some kind, although obviously removing route gaps would improve the aesthetic. If your favorite stop does not show up on the map, odds are that it was reached by a bus (and so is an unlabeled yellow dot).
I originally created a limited prototype for these maps in 2009, and expanded them for the MBTA mapping challenge in May, 2013. Thanks to the MBTA for having the competition. I am indebted to my lovely wife, Claire, and to the numerous friends and family who have shared their thoughts and encouragement, especially my brother Dan, Craig, and the folks at the Boston Python User Group.
Please browse the set of maps I have posted online, which are available by following the links through the flags in the map below. Enjoy!
These maps represent averages and estimates derived from official MBTA scheduling data. They visualize travel times in general, but do not predict a rider experience for any particular trip or time.
The source code is available under a permissive MIT license via GitHub; click here.
The maps, like all of the original content on transitboston.com, are freely available through the Creative Commons attribution license. The fine print for the contest concerning ownership was more onerous, but MassDOT kindly confirmed that the license was acceptable.