Short description
The objective of the project was to assess the feasibility to deploy dedicated service patrol vehicles to clear the arterial crash site to reestablish smooth traffic flow. The objective was to identify crash hotspots and explore the correlation site clearance time and other factors such as weather, severity of accident, etc. Here, I only focused on the hotspot analysis.
Data
The data contained 26,861 weekday daytime crashes during 3 year period for a region in Orlando shown below. The data contained
- Crash location
- The detailed response of agencies to clear the site which include
- Crash time
The first task was to explore the data
to look for hotspots
. The spatial analysis was performed in R mainly using packages sf, sp, tmap, tmaptools and osmdata.
Why R and not ArcGIS?
This was the first time, I had to perform spatial analysis in a project. I chose R over ArcGIS for two reasons:
- Reproducibility – sharing scripts underlying the analysis will allow others to build-on your work
- Flexibility – ability to write custom functions to perform analysis which would be difficult in ArcGIS. The second part of this project used this flexibility to built a numerical simulation (please click here for details)
Resources to learn spatial analysis in R
I learnt spatial analysis mainly from these two resources:
- Geocomputation with R (open-source)
- Data camp courses (paid subscription)
Siddhartha Gulhare (site should be fully functional after April 14) 