Inspiration
404 Error: answer not found
What it does
The Car-kill-ator allows the user to easily input their commute to and from two separate locations. Our website then imports the current day, time, and speed-limit to predict the likelihood that a driver will crash on their designated path. This overall percentage is composed of several percentages that describe the crash chances for smaller road segments. These segments are displayed on a map, and list theoretical distance traveled and sped on the roads.
How we built it
- found appropriate data sets
- created a model to determine the risk factor of driving on road segment using machine learning
- designed a webpage that inputs a google maps trip to predict the crash probability of a driver
- use multiple APIs from Google and other sites to obtain road information necessary for the calculations
- use server requests calling as well as python function calling to obtain necessary data
Challenges we ran into
- how to choose what data to use
- Information within databases was inconsistent
- time-constraint of the competition lead to abandonment of certain parameters (weather most prominently)
- limited prior knowledge with many of the needed language
- consistent color gradient to probability depiction
What we learned
- Hackathon is pretty cool
- It is important to make a detailed and structured plan from the beginning
- Have a basic understanding of multiple programming languages is necessary
- Asking for mentorship could have been beneficial
- Our understanding of programming languages is not enough
What's next for Car Crash Probability Calculator
- Use of more parameters to more accurately predict the crash probability
- Possibly include more data set to allow more calculation
Log in or sign up for Devpost to join the conversation.