[an error occurred while processing this directive] [an error occurred while processing this directive]
ArcNews Online
 

Winter 2005/2006
Search ArcNews
 

E-mail to a Friend
Online Only Article

In Salt Lake City, Utah, Traffic Accident History Model Built with GIS

Looking out over Salt Lake City's skyline toward the snowy and rugged Wasatch mountains, or past the chiseled stone city administration building, or toward the historic Mormon temple, observers see that the city streets are arrayed as a grid. Having been designed by Utah's pioneer settlers to be wide enough to turn around an entire team of horses, today the few horses on the streets pull coaches of skiers and tourists and operate no longer as large teams.

  click to enlarge
The bull's-eye map.

One of the many roles of Salt Lake City's Transportation Division is to study the relative safety or danger of roadways and to mitigate any situations that may contribute to accidents.

At the Transportation Division, questions often come from citizens about accidents in, and safety of, an intersection with which they are familiar. Division employees need to look at whether the mentioned intersection actually has a problem and how it compares with other intersections in the city. It's important to know what data the city has on hand and how quickly it can be accessed.

With accident data, there may be huge variation over both space and time, and being able to generalize discrete events into useful information can be quite a challenge. In Salt Lake City, filing cabinets filled with accident reports and traffic studies crowd storage rooms at city offices. Sometimes there is no simple answer to a traffic-related question.

Salt Lake City has been using Esri software for more than 10 years, and most recently, ArcGIS Desktop (ArcInfo) has been used at Salt Lake City's Transportation Division for many purposes—from planning the 2002 Winter Olympics and creating transportation master plans to organizing and maintaining data. The return on investment includes a robust method to hold and query all the division's data.

As more people use the roadways, more accidents are likely to occur. To answer a question about a particular intersection, a simple statement of accident frequency may not be adequate. The frequency, normalized by the usage, would be a more appropriate response. This usage, or the volume of traffic, comes from traffic studies that often consist of gathering traffic volumes and speeds from hoses nailed down across traffic lanes that count the passing vehicles. After the raw data is run through a program to help ensure the data's integrity, it's fed into an extensive geoprocessing model built with the ArcGIS ModelBuilder application that is included in ArcGIS Desktop.

The true benefit of ArcGIS ModelBuilder is the automation of workflows that would normally be repetitive and time consuming. Salt Lake City's traffic study model links the study data to a point on the map that indicates the location, as well as prepares many reports. As the data is referenced on the map, the Transportation Division has one piece of the puzzle needed to get an objective idea about the safety of any intersection. The police also collect data on the time and location of accidents, and these datasets are added to the map with all its associated data via the ModelBuilder application. Now, because these systems are in place and kept up to date, the stage is set to leverage both datasets against each other to derive data that is more meaningful than either dataset is on its own.

Utilizing ArcInfo, the GIS technician at the Transportation Division uses industry-standard equations to show, for each major intersection, the number of accidents that occur per million entering vehicles (N PMEV) for each year in the database. The technician then takes the average rate per year and derives the standard deviation. This same process is also done for average accident severity.

Advanced Label Expressions, a relatively new feature of the ArcGIS Desktop software, allows users to exploit scripting with branching logic and control over variables to build intelligently formatted labels for the features in GIS databases. One of these labels, constructed using VBScript and XML, shows the N PMEV for each intersection and the average severity per year. Each year's accident rate and average severity text is colored red or green if it has gotten one standard deviation better or worse than the intersection's preceding year's figure. This label is then overlaid onto bar graphs that add a visual component to the textual rates, and a bull's-eye graphic shows the recent year's accident rate as the outer ring, with the next inner ring showing the historically averaged rate, and the two internal rings showing the same for the recent year's average severity versus the historically averaged accident severity in the center.

Becoming familiar with the icon allows a city employee to visually extract accident rates and severity at a glance. To illustrate, all the rings are color coded as one standard deviation, with a yellow ring being average (-0.5 to 0.5 standard deviation), an orange ring being one standard deviation worse, and red being one worse than that, while green signifies a departure that is better than the norm.

By just glancing at the bull's-eye graphic map that has been mounted on the wall, the user has data readily available to help make decisions about individual intersections, routes across the city, or entire districts. Combining GIS datasets and rendering them in ways that provide a visual component for extensive numeric data allow Salt Lake City to monitor intersections and respond quickly to people seeking information about any major intersection in their community.

For more information, contact Kevin Bell, GIS technician II, Salt Lake City Transportation Division (tel.: 801-535-7131, e-mail: kevin.bell@slcgov.com).

[an error occurred while processing this directive]