Comparative Analysis of Intersection Data Derived from INRIX and GRIDMART

With an increase in transportation funding, the need arises to implement analytic tools to measure the effects of improvements and guide decisions regarding the effective planning and design, as well as the operation of streets. Recent advances in intelligent transportation systems (ITS) and data collection made available high volumes of data from multiple data sources. However, a significant challenge remains in digesting and understanding the large and complex data available. In the specific case of intersections, the estimation of performance metrics is challenging due to the presence of different traffic control systems, multiple data sources and multimodal interaction. The principal objective of this research is to evaluate intersection performance metrics calculated using two different data sources and using Austin, Texas, as a case study. The evaluation is based on the fusion of INRIX data and GRIDSMART data. The main contributions include development of comparative intersection performance metrics combining two data sources, implementation of all required data workflows, and analysis of the fusion of GRIDSMART data and INRIX data. The analyses suggest that while total volume estimates provided by INRIX are not likely to be accurate, and discrepancies with GRIDSMART data are larger than 20% in most analyzed cases, differences in turning movement ratios yielded more promising results with differences between the two data sources generally less than 10%. Those results provide information that can help practitioners select an appropriate methodology for evaluating intersection performance.