Quantifying Impact of Weather Condition on Travel Time

Abstract

Most transportation systems operate at capacity. Minor changes in the system could result in congestion and delays. One of the many impacting factors of transportation is weather condition. Weather conditions might lead to a totally different setting for management of transportation systems. Since weather is predictable, being able to measure the impact of weather conditions on transportation systems would help in better transportation management. Estimating dependency of travel time on weather condition will enable us to predict more accurate travel time. But it is possible that not all components of weather impact travel time equally. There are several other factors associated with travel time that interact with weather conditions to affect travel time. Other questions raising from this are: 1. Which weather component impacts travel time the most? 2. Is the impact of weather on travel time a function of time? The exercise investigates regression models to understand the effect of weather condition, accidents, and time on travel duration. Based on the identified factors parametric and non-parametric classifiers are implemented to provide class-based predictions. Lastly, the machine learning models are the rated based on accuracy, precision, recall, and Cohen Kappa score, and envisioned for various use cases.