The changing accuracy of traffic forecasts.

Area of Science:

• Transportation Engineering
• Urban Planning
• Data Science

Background:

• Despite advancements in travel demand forecasting methods, their accuracy remains under-researched due to data limitations.
• Understanding forecast accuracy is crucial for effective transportation planning and policy-making.

Purpose of the Study:

• To compile the largest known database of traffic forecast accuracy for road projects.
• To compare measured traffic volumes with forecast volumes and identify factors influencing accuracy.
• To provide insights into the reliability of traffic forecasts for planners and policymakers.

Main Methods:

• Compiled a database of 1291 road projects in the US and Europe, including forecast traffic, post-opening counts, and project attributes.
• Performed statistical comparisons between measured and forecast traffic volumes.
• Analyzed the correlation between project characteristics, economic factors (unemployment), and forecast accuracy.

Main Results:

• Measured traffic was, on average, 6% lower than forecast, with a 17% mean absolute deviation.
• Higher accuracy was associated with higher volume roads, higher functional classes, shorter time spans, and the use of travel models.
• Forecast accuracy improved over time, with recent forecasts being more accurate than older ones. Economic conditions, like unemployment, also impacted accuracy.

Conclusions:

• Traffic forecasts should be considered a range of possible outcomes, not a single definitive prediction.
• Planners and policymakers can use this research to better understand the confidence levels associated with traffic forecasts.
• Future research should continue to monitor forecast accuracy and refine forecasting methodologies.