Questions

  1. What locations in Edmonton have the most traffic collisions?
  2. What is the most common reason for traffic collisions?
  3. Are there more injuries or property damage collisions?

Dataset

We will be using a dataset provided by the City of Edmonton:

List of Top 100 Mid-block Collision Locations in Edmonton for the years 2012-2013. This file contains information about those mid-blocks in Edmonton , with a breakdown of collisions by the standard list of causes used in reporting. Collision information is provided by the Motor Vehicle Collision Information System (MVCIS), maintained by the City of Edmonton Office of Traffic Safety. This system tracks police-reported collisions involving at least one motor vehicle that occur on public roadways and which result in property damage of $1,000 or more, as well as any collisions resulting in injury or fatality. This file is meant as a companion piece to the Annual Collision Report, produced by the City of Edmonton Office of Traffic Safety. Please refer to the full report for more information on collisions in Edmonton, and definitions of the collision causes and injury types used in this file.

Exploratory Analysis

Let’s start by taking a brief look at the dataset.

There are 5 different variables (exluding the Latitude and Longitude) available in this dataset. There are 229 different locations where a collision has taken place in this dataset. Also, there are 3 unique collision types described: “Fatal”, “Injury”, and “Property”. Lastly, there are 24 different causes of collisions defined in the dataset.

Let’s try to answer the first question: “What locations in Edmonton have the most traffic collisions?” Below is a table of the 10 most common locations for traffic collisions:

Top 10 Highest Collision Locations (2013-2014)
Location Total Collision Count
HIGH LEVEL BRIDGE 107
WHITEMUD DRIVE WEST OF 122 STREET NW 106
WHITEMUD DRIVE WEST OF 66 STREET NW 68
CALGARY TRAIL NORTH OF 34 AVENUE NW 55
YELLOWHEAD TRAIL WEST OF 107 STREET NW 53
170 STREET NW NORTH OF 95 AVENUE NW 50
WHITEMUD DRIVE WEST OF 50 STREET NW 49
WHITEMUD DRIVE WEST OF 91 STREET NW 44
WHITEMUD DRIVE EB/WB EAST OF 149 STREET NW 43
WHITEMUD DRIVE WEST OF 34 STREET NW 43

Unsuprisingly, the High Level Bridge has the most collisions. For those of you unfamiliar with the High Level Bridge, take a look at the location:

Following in second highest number of collisions seem to be various junction points throughtout the Whitemud.

Now that we have examined the top locations for collisions, let’s examine the top collision reasons:

Top 10 Reasons for Collisions
Collision Cause Total Collision Count
Followed too Closely 1537
Changed Lanes Improperly 809
Ran Off Road 551
Failed to Yield to Right-of-Way 114
Left Turn Crossing Path 91
Stop Sign Violation 85
Struck Parked Vehicle 77
Improper Turn 69
Animal Action 47
Other 44
Total Number of Collisions based on Collision Type
Collision Type Total Collision Count
Property Damage 3077
Injury 481
Fatal 6
Top 5 Collision Causes Grouped by Collision Type
Collision Type Collision Cause Total Collision Count
Property Damage Followed too Closely 1295
Property Damage Changed Lanes Improperly 759
Property Damage Ran Off Road 461
Injury Followed too Closely 240
Property Damage Failed to Yield to Right-of-Way 100

Just out of curiosity, let’s see what the 5 fatality collisions look like:

Year Location Description Collision Type Collision Cause Number of Collisions
2012 BEVERLY BRIDGE Fatal Changed Lanes Improperly 1
2013 ANTHONY HENDAY DRIVE SB SOUTH OF 87 AVENUE NW Fatal Followed too Closely 1
2012 ANTHONY HENDAY DRIVE EB EAST OF 91 STREET NW Fatal Followed too Closely 1
2012 101 STREET SW NORTH OF 30 AVENUE SW Fatal Ran Off Road 1
2012 GROAT ROAD NB/SB SOUTH OF STONY PLAIN ROAD Fatal Ran Off Road 1
2012 82 AVENUE NW WEST OF 101 STREET NW Fatal Struck Parked Vehicle 1

Finally, we will visualize the top 50 locations of traffic accidents ignoring all other factors:

Further Study

There is one important factor we are unable to take into account in this analysis. We do not know what time of year the collisions took place. From the data, we only know what year the collision took place, not what month, day or hour. It would also be interesting to see the correlation of number of traffic collisions with the traffic volume at a particular location.

References