Network Screening - Sliding Window helps you rank and prioritize roadways for potential crash reduction efforts.
The Sliding Window applications in Network Screening and Advanced Network Screening break roadways into fixed-length windows for analysis and ranking.
A window is a fixed roadway segment length used for analysis. The Sliding Window application divides roadways into windows of a specified length. Both Advanced and Regular Network Screening create the windows using the same methodology outlined below.
The applications generate windows by moving a fixed-length window along each roadway in 0.1-mile increments. At each increment, the system calculates the selected performance measure for all crashes that fall within that window. Once complete, all windows are ranked according to the resulting performance measure values.
Window Examples
The following examples illustrate how window length and performance measure selection affect the generated windows and their resulting rankings along the same roadway. Note: The window locations differ even when the same roadway and crashes are analyzed.
(Example 1) Performance Measure: Crash Count, Window Length: .3 miles
In this example, windows are generated by moving a .3-mile window along each roadway in 0.1-mile increments. At each increment, the system counts the total number of crashes that fall within that window. Once complete, all windows are ranked according to the selected performance measure.
(Example 2) Performance Measure: ePDO, Window Length: .3 miles
In this example, windows are generated by moving a .3-mile window along each roadway in 0.1-mile increments. At each increment, the system calculates the ePDO for the crashes that fall within that window. Once complete, all windows are ranked by highest ePDO.
This example uses the same roadway as Example 1, but the windows are ranked by ePDO rather than crash frequency. Because the ranking measure has changed, the highest-ranked window is not necessarily the one containing the greatest number of crashes. In example 1, the top-ranked window begins at the 0.2-mile point because it captures the highest crash count. When ranked by ePDO, the top-ranked window begins at the 0.1-mile point because it captures crashes with the highest combined severity. As a result, the window locations may differ even when the same roadway and crashes are analyzed.
Note: EPDO was calculated by assigning weights to crashes based on severity and summing the weighted values within each analysis window.
(Example 3) Performance Measure: Crash Count, Window Length: 1 mile - Window Size Exception
In this example, windows are generated by moving a 1-mile window along each roadway in 0.1-mile increments. At each increment, the system counts the total number of crashes that fall within that window. Once complete, all windows are ranked according to the highest number of crashes.
In this example, the beginning mile point for the first window and the ending mile point for the last window are outside the total length of the roadway. The Sliding Window application allows this exception so that crashes occurring near route boundaries are still included in the analysis, even when the full-length window cannot be generated. As a result, some windows may be shorter than the specified window length. This allows roadway sections with crashes to remain in the analysis, even if they are shorter than the specified window length.
