November 4, 2016

Bus Crowding: Introduction

How the measure for bus crowding was decided and how it is calculated

The Office of Performance Management and Innovation and MIT researchers have developed a model to measure crowding on MBTA buses in time and space. We are building tools to analyze where and when on the street network crowding is taking place, different temporal patterns of crowding by route, and the causes of crowding. This analysis will help the MBTA determine the appropriate tools to use to address the problems.

How Crowded is Crowded?

The first step in the process is determining what constitutes crowded. As part of the revision to the Service Delivery Policy the MBTA conducted a survey that asked passengers about crowding. There are also state regulations governing bus loads.We recognize that there are different expectations in peak hours and in off-peak hours. In our definition of crowded during peak hours all passengers are counted as crowded if the bus is at 140% of seated capacity. We know even if you have a seat if that many people are in the aisle it feels crowded. Most of our 40-foot buses have 39 seats, so a peak bus becomes crowded at 55 passengers. During off-peak we count standees above 125% of seated capacity as crowded and everyone as crowded above 140%.We wanted to create a measure that took account of how long the crowding lasts. The previous measure of bus crowding just looked at the peak load on the route, but being on a bus that is crowded for one stop is different than crowded for extended periods of time. Our new model powered by our Origin-Destination-Transfer model allows us to estimate the load for the whole bus route and therefore determine minutes of crowded passenger time.

Overhead photo from inside a bus. Every seat is occupied and there are people standing shoulder-to-shoulder in the aisles. — The above image shows a bus at 150% of seated capacity.

Bus Crowding Measure

The MBTA's bus routes serve varied areas and purposes in the T's network. Because of this, routes are crowded at different times of day, and crowding can occur on both frequent and less frequent bus routes. In order to understand crowding on different types of bus routes we decided to measure both the proportion of crowded passenger minutes and the quantity of crowded minutes.We judge a route on what percent of the passenger minutes on that route are comfortable (not crowded). Less than 95% is considered to be failing our proposed standard. Data from all weekdays in the Fall 2015 rating found that 94% of passenger minutes on MBTA buses are comfortable. We have 30 bus routes that are clearly below the 95% standard. However, the quantity (meaning the number of crowded passenger minutes) is not the same on all crowded routes.

Charting showing levels of crowding through the down on three routes. Route 9 has sharp peaks inbound, Route 66 has constant low levels of crowding, and Route 111 has peaks AM inbound as well as constant low-level crowding in both direction over the whole day.

This chart was presented as part of a presentation on the Service Delivery Policy to the FMCB. We have built a data visualization tool that shows the number of crowded passenger minutes in each 30 minute segment of the service day for every bus route. This tool allows us to identify different patterns of crowding.Routes like the number 9 have a very high peak of crowding during the morning rush hour. Routes like the 66 have lower levels of crowding, but are crowded for most of the service day. The 111 has both peaks of crowding during the morning rush hour as well as lower levels for the entire service day.

Next Steps

There are different causes of crowding and different tools the MBTA and our municipal partners can use to address it. Given the many dimensions of bus crowding future posts will discuss our work to visualize bus crowding both spatially and by cause.