Dimitris Nikolaou has successfully defended his PhD dissertation titled: Machine learning-based road crash risk assessment fusing infrastructure, traffic and driver behaviour data, under the supervisoon of NTUA Prof. George Yannis. Τwo distinct databases were developed; the first one concerned motorway segments and included road crash, traffic, road geometry and driver behaviour data (OSeven telematics), while the second database concerned urban and interurban road segments of a broader area. The results revealed that crash frequency on motorway segments is positively correlated with the traffic volume, the segment length, the number of harsh accelerations and the number of harsh brakings per segment trips. Furthermore, it was concluded that harsh brakings can serve as a valid subcategory of Surrogate Safety Measures under naturalistic driving conditions, which can be used as the dependent variable either in proactive road safety analyses or in cases where detailed crash data are unavailable. It was found that harsh brakings were positively correlated with road segment length, number of trips per segment, speeding and mobile phone use.