Preliminary Estimates Model Expands its Portfolio

In early 2025, the Bureau of Transportation Statistics (BTS) began releasing preliminary estimates (PE) for monthly enplanements. The PE model, which calculates correlations between early indicators and complete historical datasets, allowed BTS to include an additional level of forecast and analysis to the monthly enplanement data, providing more timely data points for industry and partners with prompt estimates 3-12 months in advance of final audited counts.

Over the last year, BTS expanded this specialized regression-based model to forecast preliminary estimates not just for aviation enplanements, but for additional key transportation metrics: Transit Ridership, Amtrak Ridership, TransBorder Freight, and Vehicle Miles Traveled.

Aviation Enplanements

• Lead Indicator: Transportation Security Administration (TSA) checkpoint screening data (available daily).
• Target Dataset: Monthly T-100 market data (full, domestic, and international), which carries a 3- to 4-month publishing lag.
• Methodology: BTS uses a simple linear regression model incorporating seasonality controls. TSA checkpoint numbers serve as an immediate proxy for passenger volume, yielding high-confidence early estimates.

Amtrak Ridership

• Lead Indicator: Station-level ridership figures released annually by Amtrak.
• Target Dataset: System-wide annual ridership.
• Methodology: BTS applies predictive linear regression to historical indicator trends, delivering early estimates well ahead of final audited annual figures.

TransBorder Freight Data

• Lead Indicator: U.S. Census Bureau FT900 report (U.S. International Trade in Goods and Services).
• Target Dataset: Detailed U.S.-Canada and U.S.-Mexico freight flow statistics, which carry a 2-month processing lag.
• Methodology: Linear regression models bridge raw trade figures and actionable freight metrics to output preliminary monthly trade flows.

Transit Ridership

• Lead Indicator: Unlinked Passenger Trips (UPT) from the Federal Transit Administration's (FTA) National Transit Database (NTD).
• Target Dataset: Monthly national transit ridership totals.
• Methodology: Predictive regression models process monthly FTA feeds across major transit modes, including Bus, Ferry, and Rail (Heavy, Light, and Commuter Rail).

Vehicle Miles Traveled (VMT)

• Lead Indicator: FHWA Traffic Volume Trends (TVT) monthly reports, which collect hourly count data from roughly 5,000 continuous monitoring stations nationwide.
• Target Dataset: Long-term national VMT trends.
• Methodology: Developed jointly by BTS and the Federal Highway Administration (FHWA), this model integrates high-frequency traffic counts with macro-indicators to estimate total road usage.