Car accidents are a crucial issue currently present in our world today. Road crashes are the leading cause of death for people under 30. While poor driving is an issue, this can be largely attributed to poorly designed roads and infrastructure.
Significant factors that affect where crashes occur include both static and continous factors. Static features include the sizes of roads, speed limits, road curvature, and billboards. Continous features include traffic congestion and the weather.
It is important to be able to track which conditions make a road dangerous and where potential hotspots for crashes are in order to help city planners adjust in the future. Knowing what is causing hotspots is also important as it allows for smarter design.
People should also be able to know which roads and paths are the safest so that they can prepare ahead of time before driving. This ultimately promotes smarter travel. Allowing residents to quickly report accidents would also improve data collection.
Using ML, we are able to detect crash hotspots by taking in features such as the size of the road, stop signs, and billboards. We also look at constantly changing features such as weather and concentration of cars. This outputs coordinates on the roads, and we render a graphical heatmap, making it simple for residents and city officials to understand where hotspots are.
Governments (both city and state) also are able to use this to upload their city data in order to help them discover collision hotspots and figure out how to adjust for the future. We will use their data to generate heatmaps for them, allowing them to better understand the problems in their road infrastructure.
Users can view where in their city the hotspots are, allowing them to adjust before driving. This gives drivers a peace of mind, and allows them to make safer and more informed decisions. This is especially key for new drivers who don’t have a strong understanding of their roads. They can also help governments collect data by reporting crashes which gets processed in our database, overall helping the data collection process, and adding to our heatmaps.
Cities should use our process because currently, roads are very dangerous, and there isn’t an accurate and efficient process to figure out where roads are risky and how to fix them. Local governments and city planners can use our software to improve their roads in a unique way. They are provided with a heatmap of where crashes in their city can occur and how to address those issues.
We will receive data for our heatmaps from those who witnessed crashes as well as state/local governments. We have found that many cities already collect this data, but won’t release it publicly. Additionally, we would need cities to collect more data, and we have a system to allow residents to report crashes, improving our data collection process.
Residents can find a map with the hotspots of where crashes are likely. These heatmaps change based on an hourly basis and time of year to account for rush hours and temperature/weather. The common pedestrian or driver can also help improve the efficiency of our model by inputting data about crashes in their neighborhoods by interactively placing pins on the map, which we aggregate with already provided data.
Generating heatmaps relies on having both static and updated data. This includes where crashes occur as well as factors such as weather, road curvature, where signs are located, and more. Our model uses supervised learning to then generate coordinates on where it is most risky, which we then render using GeoPandas.