Using data from the Washington Metropolitan Area Transit Authority (WMATA), students in SDAL’s Data Science for the Public Good Program diagnosed some of the determinants that contribute to bus fare evasion, an issue that represents up to $20 million in lost revenue each year.

washington metro area bus


WMATA Metrobus is the sixth busiest bus agency in the United States, with a fleet of more than 1,500 buses operating on 325 routes. Metrobus provides more than 400,000 trips each weekday serving 11,500 bus stops across the WMATA jurisdiction.

Fare evasion is a national problem that affects all modes of public transport, including light and heavy rail, metro, and bus. In the D.C. Metropolitan Area, WMATA estimates that they lose approximately 10-20 million dollars a year due to bus fare evasion. This leads to revenue loss which can, in turn, result in higher fares, service cutbacks, and less revenue for maintenance and replacing older buses.

With the goal of averting further revenue loss, WMATA partnered with the Social and Decision Analytics Lab (SDAL), part of the Biocomplexity Institute of Virginia Tech, to provide insights into the problem of bus fare evasion that could be used to guide effective interventions. 


The SDAL team used WMATA administrative data, Approximate Fare Counter (APC) and Automated Fare Collection (AFC), from one work week in May 2017 to approximate where fare evaders live. The assumption was made that riders in the morning commute would correlate with a rider’s census block group.

a map displaying levels of economic vulnerability in the DC metro area

The American Community Survey was used to tell the story at the census block group level by constructing an economic vulnerability composite index. This consisted of a percentage of households in poverty (Federal), percentage of households with no vehicle, percentage of households qualifying for SNAP, and percentage of households with housing burden > 50%.  

Fare evasion is documented in the AFC data by the bus driver who presses the #8 key on the bus fare box whenever a rider boards the bus without paying. The fare evasion counts were tallied for each bus stop and then aggregated by census block group.  


The results of the team’s research found two census block groups were consistently outliers for their high number of fare evaders (identified in orange) throughout a 24-hour period.

incidences of fare evasion mapped across morning commute routes

Based on the team’s exploratory analyses, the observation was that made that not all economically vulnerable census block groups have high numbers of fare evaders, but all census block groups with a high number of fare evaders are economically vulnerable.

By aligning fare evasion with census block group and providing an indicator of economic status, WMATA can better align fare evasion interventions toward areas of higher risk.

Back to top