Big data fusion to estimate driving adoption behavior and urban fuel consumption

Abstract

Data from mobile phones is constantly increasing in accuracy, quantity, and ubiquity. Methods that utilize such data in the field of transportation demand forecasting have been proposed and represent a welcome addition. We propose a framework that uses the resulting travel demand and computes fuel consumption. The model is calibrated for application on any range of car fuel efficiency and combined with other sources of data to produce urban fuel consumption estimates for the city of Riyadh as an application. Targeted traffic congestion reduction strategies are compared to random traffic reduction and the results indicate a factor of 2 improvement on fuel savings. Moreover, an agent-based innovation adoption model is used with a network of women from Call Detail Records to simulate the time at which women may adopt driving after the ban on females driving is lifted in Saudi Arabia. The resulting adoption rates are combined with fuel costs from simulating empty driver trips to forecast the fuel savings potential of such a historic policy change.