This mobile application is a project of our H417 team for the Hack HPI 2024 hackathon on smart cities. It is a navigator that suggests the safest auto routes instead of the fastest routes. The route hazard measurement is based on historical data of accidents that have occurred since 2016 in New York City, metadata like population density and area utilization, and continuously updated data from NYC's open traffic cameras. For processing historical and metadata, we relied on the AccIndex project by Anand Deshpande, Dr. Onur Kerimoglu and Jan Meyer. To collect and process traffic camera data we used Starwit Awareness Engine. During the hackathon we managed to train the Pytorch model to epoch 20 and to a loss of 0.2 and put it partially to use in a Flutter mobile app. Backend and frontend, written in C# and Dart respectively.
H417 Team members:
Eduard Sobol, [email protected], https://github.com/Edward1an
Maksym Floria, [email protected], https://github.com/TomatoTheVeg
Mykhailo Kozyrev, [email protected], https://github.com/Smikalo
Viktor Lynok, [email protected], https://github.com/Lyynnnx
Yevhenii Karpushchenko, [email protected], https://github.com/SunJet72