Behind the scenes with MBTA data.

Using data to make decision on whether a dedicated bus lane should be proposed in various areas across the greater Boston area.

The following is a guest post from Scott Hamwey, Manager of Long Range Planning at MassDOT's Office of Transportation Planning.

MassDOT and the Central Transportation Planning Staff (CTPS) undertook an analysis to help us prioritize dedicated bus lanes where they were most needed (based on high ridership and high on-street congestion) and where we could make the strongest argument for their implementation (where bus passenger trips make up a significant share of the person trips in a corridor). In this post, we explain the steps of the analysis which was conducted during the spring of 2016.

Map showing high-ridership bus corridors in the MBTA region.

This first map shows our universe of high-ridership bus corridors. We looked at ridership for each roadway segment that exceeded 1,500 weekday passengers in at least one direction. Looking at roadway segment rather than route means that places shared by several moderate ridership routes (like Washington Street in Roslindale) as well as those with one high ridership route (like route 32 on Hyde Park Ave) both made the cut. This map was generated using Fall 2015 composite automatic passenger counter (APC) data. 

Maps showing congestion-related delay on high-ridership corridors

The second map shows levels of congestion-related delay on those high ridership segments identified in the first step of the analysis. To estimate this we first identified average travel speeds on each segment for every minute of the day. These average speeds were taken from 42 days of data from CTPS’ travel speed database which was purchased from INRIX, a company which collects roadway speed data from over 250 million real-time anonymous mobile devices and vehicle fleets equipped with GPS. A comparison was made between travel speed at the time of each bus trip, and a “free flowing” speed that was the lower of either the posted speed limit or the INRIX speeds between 2am and 4am. We didn’t use actual bus speeds to compare congested with free-flowing speeds because of the challenges presented by the varying lengths of bus dwell times between these time periods. One shortcoming of using INRIX was that there was not data for all roadway segments.

The map represents congested related delay in terms of passenger hours per mile, with those segments in red having the highest levels of passenger delay and those in green having the least. At the high end is the portion of Huntington Avenue between South Huntington and Brigham Circle, where on the average day the delay is 161 passenger hours per mile in the inbound direction (and this doesn’t even include the passengers on E Line running in mixed traffic here!). Even though this stretch of Huntington is less than ½ mile long, that translates into 51 minutes of passenger delay per trip between 8am and 9am (or a total of 758 total passenger hours of delay across all 15 trips on routes 39 and 66 during that hour). 

Map showing percent of total people on the corridor made of of riders on buses

The final map focuses on just a few of the high ridership roadway segments – just those that had the highest levels of passenger hours per mile delay. Here, we’re interested in how large a percentage of roadway users are represented by these slow moving bus passengers, with red segments being those with a high share of bus riders (in this case during the AM peak hour). We used peak hour bus ridership data from APC, and peak hour traffic volumes derived from City of Boston or City of Cambridge intersection turning movement counts. After subtracting buses from the traffic counts total, each motor vehicle was assumed to carry 1.38 occupants (based on input received in the most recent Massachusetts household travel survey). The map shows those with the highest share of bus passengers in red, with Washington Street in Roslindale having the highest share – 58.4% of roadway users on buses in the northbound direction during the AM peak hour. That’s a pretty big share of bus riders even when there’s no time savings advantage to being on the bus.  

This analysis has helped us identify locations for meaningful improvements for bus passengers that could be implemented in the short term, including the addition of an inbound bus lane in the City of Boston’s new North Washington Street bridge designs. Now MassDOT and the MBTA are working with our partners at the City of Boston and other municipalities to discuss implementation.