This is a project of CSE583 by Katharine Chen, Tianqi Fang, Yutong Liu, Shuyi Yin.
Crash4viz is an application that integrates various data sources of highway crash in Washington State over the last 5 years. This tool facilitates visualization, exploration and analysis by average driver and professional users.
Our technology review introduces problem background, general use cases, data sets and initial data wrangling. It is online at technology review.
Our final presentation per request of CSE583 project requirement introduces project background, database design, use cases, design, structure and lessons learned. It is online at final pre.
The project has the following structure:
|-- crash4viz
|-- interface.py
|-- requirements.txt
|-- setup.py
|-- crash4viz
|-- mapping.py
|-- mapping_funcs.py
|-- mlpredict.py
|-- dataprep/
|-- tests
|-- test_dataprep.py
|-- test_mapping.py
|-- test_mapping_funcs.py
|-- test_ml.py
|-- test_output/
|-- data
|-- hsis-csv/
|-- crash-merged/
|-- docs
|-- pre-final pre.pdf
|-- examples/
|-- data handbook/
|-- design docs/
|-- final pre/
|-- tech review/
|-- outputs/
To install this program, run:
git clone https://github.com/syin3/crash4viz.gitcd crash4vizpython setup.py install
To launch the crash visualization interface, run (from the top-level of the github repo):
python3 interface.py
Select the environmental, road, and incident type features you would like to visualize and press Show map, the interactable html map will be output into the directory outputs.
For users interested in testing the time to completion of each type of map, you can run (from the top-level of the github repo):
python -m unittest crash4viz.tests.test_mapping
Currently this will run timed tests for each type of map using a default selection of crash data features, which may differ slightly from feature to feature.