Long Island Railroad (LIRR) is the largest commuter rail service in the United States, serving 87.6 million riders in 2015. LIRR relies on forecasts of passengers and ticket revenue for budgeting and capital planning purposes.
How we helped
Steer was requested by LIRR to develop a series of ticket sales and ridership forecasting models and forecasts.
The models were based on econometric techniques, which use statistical tools to estimate the relationships between ticket sales and economic variables (notably regional employment), as well as fares and fuel prices. The models also incorporated the effect of changing service levels and on-time performance as variables that influence ticket sales.
The model development distinguished between different ticket types offered by LIRR, including commuter (monthly, weekly and one-way peak period), mixed commuter and leisure (one-way off-peak), and a number of different leisure ticket types. Each of these ticket types were modeled separately to account for considerable difference in behavior and price responsiveness in the markets.
Before estimating the models, Steer engaged in a large exercise in database development, taking raw ticket sales data, which included millions of records, to develop the summarized data that was used as the basis for the modeling.
Successes and outcomes
Following the successful development of the individual ticket sale models, Steer also developed interactive forecasting tools based on the modeling for LIRR to develop periodic forecasts for budgeting and capital planning.