Network Screening

Network screening is a process for ranking locations on the road network according to their crash risk, from high to low. The process is sometimes referred to as blackspot identification. Examples of locations may be a section of roadway, an intersection, a freeway ramp or a midblock pedestrian crossing. Those locations identified as highest risk can then be investigated thoroughly to identify crash patterns, contributing factors, and appropriate countermeasures.

Network screening can also be used to implement a policy such as installing red-light cameras at the signalized intersections most likely to benefit from them. In such a case, intersections would be screened for crashes and identified as related to red-light-running or angle-type crashes.

Back to top

Following a formal process for prioritizing locations for study or improvements ensures that decisions are made based on data rather than subjective perceptions. The process and documentation of issues provides a transparent paper trail to justify why locations were or were not prioritized, and this can be demonstrated for the purpose of any legal proceedings.

Back to top

There are several methods available for Network Screening. Which approach should be applied depends on what data are available and the technical abilities of the persons conducting the screening.

All the methods are based on the use of crash data as a minimum, but they differ in terms of reliability. Generally, these methods can be categorized into three groups:

1. Methods that look at simple counts of crashes or crash rates
2. Methods that compare crash counts or rates to what are normal counts or rates at other similar locations
3. Sophisticated methods that use Safety Performance Functions (SPFs) to compare a location’s long-term expected crash count to that for similar locations

As an illustration of one approach to Network Screening, consider the data in the table below. The table shows five ranked intersections. The data provided includes the observed number of crashes in the past three years and the average annual daily traffic at each intersection. Also shown is the predicted number of crashes per year from an appropriate intersection SPF and the long-term expected crash frequency per year that is a weighted average of the observed and predicted. The intersections have been ranked in descending order by the excess crashes per year which is taken as the long-term expected frequency per year minus the SPF prediction of crashes per year. In effect, this is an estimate of how many more crashes are expected to occur at each intersection compared to what is normal for similar locations.

Illustration of Network Screening of Intersections

Detailed information on the available methods and the advantages and disadvantages of each can be found in the following documents:

1. Transportation Association of Canada: Guidelines for the Network Screening of Collision-Prone Locations
2. AASHTO’s Highway Safety Manual

Back to top

Data requirements depend on the selected methodology. Crash data for locations to be screened are always required. Other required data may include:

• Traffic volumes for the same years as the crash data
• Geometric data sufficient for classifying facility types (e.g., urban four-legged two-way stop-controlled intersection)
• SPFs and any geometric or other location-specific information required to apply them

Back to top

Required skills will depend somewhat on the methodology applied. For more advanced methods, the ability to interpret statistical analysis results is required. If specific tools to implement the procedures are developed in-house then software programming skills may be needed.

Back to top

While specific tools for Network Screening are not necessarily needed, there are some available. This list should not be considered exhaustive.

1. Transportation Association of Canada: Guidelines for the Network Screening of Collision-Prone Locations
The guidelines provide detailed information on Network Screening methodologies that can be used by jurisdictions with a range of data and technical capabilities.

2. AASHTO’s Highway Safety Manual
AASHTO’s Highway Safety Manual (HSM) includes a chapter on Network Screening that provides detail on several methodologies that can be applied.

3. DiExSys Vision Zero Suite
DiExSys Vision Zero Suite (VZS) is a software that provides predictive, diagnostic and analysis tools which reveal the nature and magnitude of the safety problems on the network, segment and intersection levels. VZS is capable of evaluating cost-effectiveness of safety improvement strategies, enabling network screening, producing observational before-and-after studies, leveraging GIS mapping, and recognizing infrastructure and behavioral patterns, facilitating implementation of countermeasure deployment and targeted enforcement.

4. TES Software
TES is a software for transportation data management and analytics that includes Network Screening applications.

5. AASHTOWare Safety
AASHTOWare Safety is a Software as a Service (SaaS) platform specifically designed to meet the unique needs of state and local transportation agencies in the area of highway traffic safety management. The integrated Safety Data Warehouse that the platform is built upon, houses all the necessary data and supports the six steps in the Roadway Highway Safety Management Process: 1) Network Screening, 2) Diagnosis, 3) Countermeasure Selection, 4) Economic Appraisal, 5) Project Prioritization, and 6) Safety Effectiveness Evaluation.

Back to top

Back to top