Fatality Analysis Reporting System
- 1. FARS Data Set Analysis and Visualization Author: Karthik Ranganathan
- 2. What is it? Contains Accident Information on • Date • Weather • Person • Vehicle • Location The program collects data for analysis of traffic safety crashes to identify problems, and evaluate countermeasures leading to reducing injuries and property damage resulting from motor vehicle crashes. Author: Karthik Ranganathan
- 3. Methods and Work • Cleaning and Wrangling of more than 100 coded elements • 600 page catalogue to decode • Understanding the method of data collection • Finding the key elements for Exploratory Data Analysis • Visualizing the right information Author: Karthik Ranganathan
- 4. Raw Data Author: Karthik Ranganathan
- 5. Cleaned Data Author: Karthik Ranganathan
- 6. Findings • The data gives a lot of insight into multiple variables playing a role in the crash • By visualizing more than 50 factors, here are some of the most impactful insights Author: Karthik Ranganathan
- 7. Age • Overs the years trends have been showing increase in accident rate in age group of 65 years and above • Aging population? Author: Karthik Ranganathan
- 8. Accidents due to Drinking Note: The high rate of accident on Saturdays and Sundays but surprisingly in the non drinking category Author: Karthik Ranganathan
- 9. Initial Impact • An unlikely inference can be made from this data • We can see that sitting on the rear right side of the car will be most safest for initial impacts (given that the seat belt is fastened) • This is purely a inference and many other data points needs to be collected in order to verify and prove the claim Author: Karthik Ranganathan
- 10. How do the States fair? Author: Karthik Ranganathan
- 11. Better View Author: Karthik Ranganathan
- 12. California, Florida and Texas Author: Karthik Ranganathan
- 13. Deaths due to road accidents per state Is it telling us everything? Total 2014 2015 Percent Change Traffic Fatalities 32,744 35,485 7.72 Population (x1000) 318,856 321,418 0.80 Licensed Drivers (x1000) 214,093 218,086 1.87 Registered Vehicles (x1000) 260,354 263,608 1.25 Vehicle Miles Traveled (x 100 million) 3,025,655 3,095,373 2.30 Author: Karthik Ranganathan
- 14. Fatalities per 100,000 licensed drivers Author: Karthik Ranganathan
- 15. Fatalities per 100 million miles travelled Rhode Island and Massachusetts have the most safest drivers. Really?? Author: Karthik Ranganathan
- 16. Which route to take on which day? Author: Karthik Ranganathan
- 18. Weather Condition? Author: Karthik Ranganathan
- 19. Road Type? Author: Karthik Ranganathan
- 20. May be Car Model Year? Author: Karthik Ranganathan
- 21. Conclusions • Need to increase surveillance of accident prone areas by thorough analysis of the heat map • Increase in safety checks of cars which are in the age bracket of 8- 17 years • Higher police presence and installation of speed cameras in Four-Way Intersection and T- Intersection • Run driver education campaigns in Mississippi, Montana, South Carolina and, Wyoming to tackle the high rate of incidence • Also major policy changes need to be implemented in California, Florida and, Texas with respect to Speed restriction and heavier fines Author: Karthik Ranganathan
- 22. Take Away • The devil is in the details • Could have looked into more data which included injuries and other incidence reports • Every situation in unique and the categorization of data has to be more complex • Generalized lot of data to simplify the visualization • By toning the data down to specific roads which have higher accident rates we can suggest detailed plans to reduce the rate • Additional Years could give us information on various trends and accurate insights Author: Karthik Ranganathan
- 23. Thank You Author: Karthik Ranganathan