Data science project built using New York State DMV data to predict crash severity.
In this repo you will find the following files:
- My powerpoint presentation on this project going over each step of the process as well as my results. This is the best place to start.
- A jupyter notebook covering the project's feature engineering.
- A jupyter notebook where I created the feature covering each county's impaired driving rate.
- A jupyter notebook covering the project's modeling.
The data used in this project comes from Kaggle and was originally uploaded by the State of New York.
Please take a look at the dashboard I made for this project as well.
Thank you for looking through my work, let me know if you have any comments or questions.
See below for a demo of the streamlit app I created using the model trained for this project. I further put the app inside of a docker container for ease of running. The screenshots and clip are from a test container deployment on Play with Docker. This app currently lives on Streamlit Share, it can be found here.
