Along with some colleagues (Deng Bin, Vassilis Zachariadis, and Ying Jin) I have recently been working with some GPS data from taxis in Beijing. We have developed a method to utilise this data as input into urban transport models. Our paper has recently been published in the Eurpean Journal of Transport and Infrastructure Research.
The paper can be downloaded from here: link.
The collection of actual traffic delays and road traffic speed data is essential in modelling urban transport resource efficiency, congestion and carbon and pollutant emissions, which is in turn part of core empirical basis for evidence-based policy making for improving urban sustainability. This data collection has also been one of the most expensive and time-consuming tasks, which restricts how well and how often the models can be built and validated, often to the extent that urban transport models have to rely on severely outdated data with sparse coverage. New smart data such as GPS vehicle traces has raised the prospect of remedying the data shortage, but for operational and data protection reasons often only low-sampling frequency traces are available. This paper proposes a novel method for estimating actual, congested link-speeds from low- sampling frequency taxi GPS traces that are publicly available. The method is based on a path inference process and is applied over a detailed road network in a large city region. It shows that low frequency GPS trajectories can significantly improve the spatial and temporal resolutions of traffic speed data for transport modelling and policy analysis. This opens up the prospect of improving road operation performance, managing travel demand and optimizing urban circulation.