Data Visualizing NYC Traffic Collisions
The skills the author demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.
Introduction
For this project I chose to build an R Shiny App (app here - github here) to data visualize traffic collisions that occurred from 2012-2017 in the five boroughs of New York. There were over one millions accidents over this time period, with over one thousand causing fatalities. Insights gained from this app have the potential to contribute to one of Mayor Bill De Blasio's important initiatives concerning transportation within New York City, Vision Zero, whose goal is to reduce traffic fatalities to zero.
Data Set
The dataset I used comes from one of Enigma's many free (account registration required) datasets and includes data concerning date, time, location, and fatality count by collision (here). I chose to visualize the data both graphically and geographically.
Data
Users are able to visually browse the data set by day of week, hour, and borough to gain insight into traffic patterns. The differences between when accidents occur on weekends and weekdays are large.
Day of Week
The different boroughs also have substantially different accident profiles by time. Notably, the accidents in Manhattan, true to its name as the city that never sleeps, have a different distribution from the other boroughs, with more of its collisions occurring around midnight. With this increase in collisions near midnight comes a corresponding decrease in accidents during the morning and evening commutes.
Borough
Users can also visualize the fatal accidents that occurred in this five year range. This subset of the data is especially important as it can lead policy makers to both understand what similarities there may be in these accidents as well as the locations where the most lives can be saved.
Conclusion
There are numerous directions to head in for further exploration of this data set; the full data set also includes information about pedestrians, cyclists, and types of vehicles involved in and primary causes for each accident. Exploring where accidents occur and for what reason may yield valuable information for reducing accidents and loss of life as well as set up the possibility for informative experimenting and testing with local targets.