The objective of this Diploma Thesis is to examine the factors influencing speed limit violations across the entire road network of Athens. The study utilizes data collected by OSeven Telematics and OpenStreetMaps, which includes information such as road geometry indicators, safety measurements and driving behavior metrics. Statistical models and machine learning algorithms were developed for two scenarios aiming to predict speeding violations and understand the factors influencing them. Τhe first scenario focused on determining the presence or absence of speeding violations, while the second analyzed speeding relative to a specific threshold. Overall, ten models were created. Through these different models, it was possible to highlight the importance of specific factors and their ability to predict the probability of speed limit violations. The results demonstrated that speeding has a statistically significant correlation with various variables and improving driving behavior will consequently lead to a reduction of road crashes. The main factors affecting the likelihood of speeding in the examined road sections are the number of trips, the road section length and the percentage of mobile phone use, while slopes presented the least impact.
| ID | ad172 |
| Presentation | |
| Full Text | |
| Tags |
