1. Home
  2. RIP
  3. View Record
https://www.nationalacademies.org/trb/events

Improving Holiday Congestion Forecasting on Interstate Highways

Holiday travel periods present a significant challenge for speed and congestion prediction due to highly variable travel behavior, event-driven surges, and inconsistent congestion patterns. The Virginia Department of Transportation (VDOT) currently relies on historical averages of travel speeds based on INRIX TMC data from the past three years, averaged by day and week of the holiday period, to predict likely congestion levels during holidays. While this method has shown good accuracy for holidays that fall on the same day of the week and month, it is less effective for holidays that occur on specific dates, such as the 4th of July and Christmas Day, where travel patterns shift year to year based on the day of the week of the holiday. This research aims to develop an advanced framework to improve predictions of travel speeds and the expected levels of congestion during holidays designated by the Commonwealth of Virginia. The project will explore a variety of modeling techniques- including machine learning classifiers, regression models, and rule-based systems to determine the most effective method for this application. The models will leverage 30-minute interval INRIX XD speed data between 7:00 AM and 12:00 AM for holidays, on Virginia’s interstate network for 2022–2024. Additional input features will include such factors as time of day, date context (e.g., "day before holiday," “day after holiday”, day of week), direction of travel (e.g., inbound, outbound, or balanced conditions), urban/rural classification, and VDOT district. In addition to XD data, the research team will evaluate the incorporation of supplementary data sources such as roadway weather information system (RWIS) or National Oceanic and Atmospheric Administration (NOAA) weather data, incident reports, and historical volume trend overlays to enhance model performance. Non-holiday data will also be analyzed to adjust for annual baseline changes in traffic volume, supporting more robust holiday comparisons.