Repository of code/work for accident-analysis subgroup in AIEA Lab (ATV), UCSC
In this project, we are analysing autonomous vehicle (ATV) accident reports generated by the Department of Motor Vehicles, California (DMV) from the year 2016 to 2022. These reports have details about the accidents which includes (but not limited to) date, time, place, car manufacturer, location of damage on ATV, weather, and road conditions.
Major amount of the work is done using Python. Here is am list of libraries that are used until now:
- beautifulsoup
- matplotlib
- pandas
- seaborn
- pypdf
- ocrmypdf
Install required dependencies with: pip3 install -r requirements.txt
The installation guide for most of the libaries can be found on their websites. Make sure to download all the libraries required in each step so that the code runs correctly.
-
Collect DMV reports
-
Create dataset by reading the files in PDF/scanned format
-
[o] Analyse the data
-
Possible Tangent: Hazard indicator based on data (Not pursuing)
-
Possible Tangent: Visualing the accidents using openai (Not pursuing)
-
[o] Expanding to different states (using NHTSA data to create a merged dataset)
-
-
[o] Research Paper : https://www.overleaf.com/read/nnzshqsjchjv#5d237a
- Fork the Project
- Create your Feature Branch (
git checkout -b feature/name_branch) - Commit your Changes (
git commit -m 'Added some feature_name') - Push to the Branch (
git push origin feature/yourname_branch) - Open a Pull Request
Distributed under the MIT License. See LICENSE.txt for more information.
Feel free to contact any of the students mentioned below in the acknowledgements.
The project is being done under professor Leilani Gilpin (lgilpin@ucsc.edu) as a part of the Artificial Intelligence Explainability Accountability (AIEA) Lab at University of California, Santa Cruz.
Following are the students who have contributed to the project:
- Rohan Ghosalkar (Graduate | CSE | rghosalk@ucsc.edu)
- Tommy Chen (Undergraduate | CSE | tchen175@ucsc.edu)
- Tong Jia (Undergraduate | CSE | tojia@ucsc.edu)