WisDOT Crash Statistics Dashboard - Shiny Contest Submission

WisDOT Crash Statistics Dashboard

Authors: Jacci Ziebert

Abstract: This is an interactive crash dashboard made for the Wisconsin Department of Transportation. There are multiple charts separated by category, such as bike/pedestrian involvement and driver behavior. Data can be selected by county, year, crash type, crash flag and features a hexbin analysis to show crash densities.

Full Description: I work as an analyst for the Wisconsin Department of Transportation. I wanted to make a dashboard that combines the geospatial element of crashes to data visualization. The dashboard is both dynamic and responsive so the user can filter different crash flags and see how varying factors and crash locations changes. I also made filters and charts that show a variety of driver behavior and characteristics that may have contributed to a crash.

There are four years of data (2017 - 2020) and data is stored in a SQLite database. I found this was much quicker than when I had it in an fst format. I wrote a script that transforms and analyzes our crash data and puts it into the database so I can add years or make a new database very easily as 2020 data is preliminary. The GitHub does not have the crash database due to files size. It can be available upon request.

I used the light-grey theme in dashboardthemes. I also changed some elements in this theme and also added custom HTML and CSS.


Keywords: transportation, crashes, leaflet, plotly, golem, traffic, accidents,
Shiny app: https://jacciz.shinyapps.io/Shiny_Crashes/
Repo: https://github.com/jacciz/shiny_wisdot_crash_dashboard
RStudio Cloud: https://rstudio.cloud/project/2302770

Thumbnail:
image

Full image: