Disclaimer about Using Crash Data

The following is the agreement that every user must accept in order for them to use crash data. Please review it regularly.

The information contained on this web site is compiled from crash reports submitted to the Utah Department of Public Safety with post-processing by the Utah Department of Transportation. The Crash Studies Team at UDOT is committed to providing the highest quality crash data to customers. However, we cannot guarantee that all crashes are represented in the data nor can we ensure that all details pertaining to a single crash are accurate.

This website is provided for the purpose of identifying safety concerns, with the intent of funding low cost safety improvements. When using this data, always reference UDOT as the source with a disclaimer that includes “this data is protected under 23 USC 409.” Never disclose this information, or analysis that uses this data, in reply to any legal action.  UDOTcrash data is protected from discovery and shall not be admitted as evidence in any court.

Overview

Network Screening is for reviewing and analyzing safety metrics for identifying potential safety projects. The metrics are shown in conjunction with crash data details and summary level information. Below is an introductory video, which explains the Crash Query, Network Screening, and Safety Analysis applications:

For UDOT, Network Screening includes the three safety models described below.

Model #1: Safety Index

The Safety Index is a metric that has been used by UDOT for several years to identify high risk crash locations based on crash rates. The Safety Index assigns each segment a score based on the crash rate, severe crash rate, crashes per mile, and severe crashes per mile. Each rate is compared to rates across its functional class and scored accordingly. The data shown in the application is based on data from 2009-2013.

Model #2: Utah Crash Prediction Model (UCPM-Frequency Model)

The Utah Crash Prediction Model is based on an advanced statistical model that uses roadway attributes and crash history to identify which roadway segments present the greatest crash risk. Segments are ranked statewide based on the model. The top ranked segment (#1) presents the greatest crash risk statewide based on the parameters of the model. The following attributes are the key factors to consider when evaluating UCPM data:

  • Statewide Rank: This is where the segment ranks statewide, the number one segment presents the greatest crash risk.
  • Total Crashes: The total number of crashes on that segment (from 2009-2013)
  • Expected Crashes: The number of crashes we would expect to occur on that segment based on its attributes and the results of the model.
  • Percentile: The percentile tells how likely the observed outcome (total crashes) was compared to the expected outcome (expected crashes). A percentile of 50% indicates that what we observed was exactly what we expected. A percentile of 0% indicates that what we observed was much less than what we expected, and a percentile of 100% indicates that what we observed was much greater than what we expected.

Model #3: Utah Crash Severity Model (UCSM-Severity Model)

The Utah Crash Severity Model is based on an advanced statistical model that uses roadway attributes and severe crash history to identify which roadway segments present the greatest risk of severe crashes. Segments are ranked statewide based on the model. The top ranked segment (#1) presents the greatest severe crash risk statewide based on the parameters of the model. The following attributes are the key factors to consider when evaluating UCSM data:

  • Statewide Rank: This is where the segment ranks statewide, the number one segment presents the greatest crash risk.
  • Severe Crashes: The total number of severe crashes on that segment (from 2009-2013).
  • Severe Probability: The Severe Probability represents the probability that a single crash occurring on that segment results in a severe or fatal injury. A high probability means that any crash occurring on that segment is very likely to result in a severe crash, and low probability means that any crash occurring on that segment is not likely to result in a severe crash.
  • Observed Probability: The observed probability represents the probability of the observed number of severe crashes occurring. A low probability means that the observed number of severe crashes was not likely to occur, indicating that the segment presents greater risk than expected. A high probability means that the observed number of crashes was likely to occur, or was an expected condition.
  • Expected Severe Crashes: The number of severe crashes we expected to occur based on the roadway attributes and results of the model.

Filtering in Network Screening

Filtering in Network Screening filters based on segment attributes, not crash data. Because you are reviewing segment-based metrics only those segment attributes can be used in filters. Crash data is represented in the segment details window after clicking a segment.

Download

To download a .csv file (which can be opened in Excel) of crash records, simply navigate to the table view and click the "CSV" button along the bottom. The crashes included in the download will reflect any filters that are currently applied. Vehicle-level data cannot be downloaded, only crash-level data.

Did this answer your question?