Aspasia Gonidi, “Investigation of the effect of pandemic restriction measures on Greek driving behaviour during COVID-19 through deep learning techniques”, Diploma Thesis, NTUA, School of Civil Engineering, Athens, July 2021.

The aim of the present Diploma Thesis is to investigate the evolution of pandemic impact on driver behavior in Greece using multivariate time series analysis. Driver behavior and pandemic’s data were collected for Greece. For the data analysis, VAR and SVAR multivariate time-series models were developed, using pair of variables, with the first variable referring to driving characteristics and the second one to the pandemic evolution. The aforementioned methodology resulted in three models estimating driver behavior during the pandemic, from which it was demonstrated that the covid-19 cases’ increase causes increase in harsh accelerations, probably due to psychological pressure but also due to the lower traffic volumes at the road network. Furthermore, due to the significant differences in the two covid-19 “waves” in Greece, models’ predictions are valid up to September 2020.