Forecasting Car Accidents in Iceland and the US: A Machine Learning Approach

This project won 2nd place in the hackathon "Hack Reykjavik 2023" in the category "AI on the Edge". See video presentation below or here (Our presentation starts at 16:58).

Is it possible to reliably predict the amount of car accidents based on weather conditions(temperature and wind speed)?

Yes! Gunnar Þrastarson, Tomas, and I built a machine learning model that, despite the noise in the data, can predict the number of car accidents in Iceland for a given month based solely on temperature. Additionally, we developed a model that can predict the amount of car accidents in the US with 96% accuracy by considering both temperature and wind speed.

These models can be used to enhance accident prevention measures, optimize resource allocation, improve traffic management and planning, aid in insurance and risk assessment, and contribute to assessing the impact of climate change on road safety.

Read the full report below or here. If you would like to learn more about our work or explore potential collaborations, please feel free to reach out. Let's work together to make our roads safer and more secure!

It appears you don't have a PDF plugin for this browser. No biggie... you can click here to download the PDF file.