Ziakopoulos A., Karahlis N., Yannis G., “Leveraging naturalistic Connected LCV data for spatial surrogate safety measure applications” at the Proceedings of the IRF Global R2T Conference (9-12 December 2025, Los Angeles).

The aim of the present paper is to showcase both the contents and the findings of a large dataset of naturalistic connected vehicle data and to examine its potential for road safety analysis. Specifically, a large dataset of LCVs is presented, collected from the Greater London Area in the UK and their acceleration values are analysed, together with map visualizations which can support policymakers and stakeholders. The data is contrasted with LCV crashes, in an overlapping study area for the same period. Several diagnostic spatial analysis techniques are explored, such as cross-Nearest-Neighbour analysis for concentration comparison of two different datapoint patterns and Ripley’s Cross K12 function, quantifying the spatial relation of the two different spatial datapoint sets. The results reveal a significant spatial correlation between connected LCV harsh braking incidents and crashes involving LCVs. In other words, LCV crash locations are more likely to occur near harsh braking points in the dataset than would be expected from a random spatial distribution. Further research might involve detail spatial regression modelling for the data in order to discover its full potential for proactive road safety analyses and correlation with specific risk factors.