Bottlenecks Summary Report

[kengodleskidot]

The information below will help us in creating a bottlenecks summary table to be used for reporting/dashboards.

Summary screen of bottlenecks in PeMS

Description of report in PeMS: The PeMS system attempts to identify bottlenecks on the freeway system. PeMS defines a bottleneck as a specific location, at a specific 5-minute time point, when there is a large enough speed drop from one detector station to the next. PeMS runs this algorithm for each VDS and for each of three shifts, AM shift (5am-10am), noon shift (10am-3pm), and PM shift (3pm-8pm), for every day. For each bottleneck, we record the duration of the bottleneck, the distance upstream, and the total delay. The details of the bottleneck identification algorithm are given on the Calculations page of the Help section of the web site.

In this section we allow users to list the top bottlenecks that have been identified over a given time period. The user needs to select a time range that they are interested in and the specific shift and then we display a table showing the results. The table lists out any detector station where a bottleneck was identified during the select days. Users can choose to sort the list according to the number of days that a bottleneck was identified at that site.

The non-obvious columns in the results table are:

VDS: The detector station where the bottleneck was identified. This station is upstream of the actual bottleneck. In other words, the next station downstream should be downstream of the bottleneck and hence in free-flowing conditions. This column links to a plot of the delay at this detector versus the time of day for the selected time range. Name: The name for the detector as supplied by the agency. Fwy-Dir: The freeway-direction for this detector. The freeway column links to a timeseries contour plot showing speed over space and time for this freeway-direction. By default the places where PeMS identified a bottleneck are drawn on the contour plot. No. of Days Active: This is the number of days that a bottleneck was spotted at this location. Avg Extent (mi): For each day, we measure the distance upstream that the bottleneck stretched for every 5 minutes. We then take the median of these distances for the duration of the bottleneck and call that the spatial extent of the bottleneck for that day. This column is the average of those spatial extents for all of the days that the bottleneck was active. This column links to a map centered on this VDS showing the bottlenecks as differently sized dots. Avg Delay (veh-hrs): This is the average delay with respect to 60 mph that this bottleneck caused on the days that it was active. Avg Duration (min): This is the average length of the time that this bottleneck was active on the days that it was active. A couple of things to note about this feature:

A bottleneck is really any location where there is a persistent drop in speed upstream. This includes locations like merges, or large on-ramps, as well as incidents. Many incidents are caught by the bottleneck identification algorithm. You'll notice these in the results listing because they occur on only a single day but cause a lot of delay. It's usually a good idea to choose to normalize the results by the number of days that the bottleneck was active. This will then show the average delay per day that the bottleneck was active. Bottlenecks observed at freeway border locations may actually belong to locations further down stream, on the other side of the border.

[jkarpen]

@kengodleskidot When you return could you please add size/priority to this one, and sprint assignment?

[jkarpen]

May not happen in sprint 2024-16 depending on how other bottleneck-related issues go.

[jkarpen]

Work can start on this when the Bottleneck Spatial Extent work is completed, however partial dependency still exists on the Bottleneck delay issue #256 .

[jkarpen]

Marking as paused, focus is on the Speed calculation issue for now