National Academies Press: OpenBook
« Previous: Chapter 3 - State of the Practice
Page 21
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 21
Page 22
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 22
Page 23
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 23
Page 24
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 24
Page 25
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 25
Page 26
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 26
Page 27
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 27
Page 28
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 28
Page 29
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 29
Page 30
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 30
Page 31
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 31
Page 32
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 32
Page 33
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 33
Page 34
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 34
Page 35
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 35
Page 36
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 36
Page 37
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 37
Page 38
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 38
Page 39
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 39
Page 40
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 40
Page 41
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 41
Page 42
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 42
Page 43
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 43
Page 44
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 44
Page 45
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 45
Page 46
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 46
Page 47
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 47
Page 48
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 48
Page 49
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 49
Page 50
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 50
Page 51
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 51
Page 52
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 52
Page 53
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 53
Page 54
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 54
Page 55
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 55
Page 56
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 56
Page 57
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 57
Page 58
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 58
Page 59
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 59
Page 60
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 60
Page 61
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 61
Page 62
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 62
Page 63
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 63
Page 64
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 64
Page 65
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 65
Page 66
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 66
Page 67
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 67
Page 68
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 68
Page 69
Suggested Citation:"Chapter 4 - Case Examples." National Academies of Sciences, Engineering, and Medicine. 2021. Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation. Washington, DC: The National Academies Press. doi: 10.17226/26094.
×
Page 69

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

21 The following use cases were collected from in-person and phone interviews of state DOTs and their partners who responded to the survey. Nineteen use cases were documented. Twelve focused primarily on the use of probe-based speed data and are shown in Table 5. Seven of the 19 use cases focused more heavily on trips and people-movement data. These use cases are shown in Table 6 and are arranged by primary category of use (planning, operations, and pandemic response). For each use case the agency/interviewee provided answers to the following questions that have been divided up into four sections for each use case: • Overview: What was the problem that was being solved or the overall goal of the analysis? • Data Source Description: What data were used and why? • Tools and Analysis Description: How did the agency accomplish their goal using the data and tools? • Results and Outputs: Was the agency successful? What lessons learned or relevant insights can be shared? While not a part of the official survey, interviewees were asked about their knowledge of their peer agency use cases documented in this synthesis. The majority of interviewees were unaware of the activities that other agencies (and sometimes others within their own agency) were under- taking with similar data. Lastly, several interviewees provided thoughts on anticipated use cases. These have been documented in short paragraphs and can be found in Appendix C. Conducting Incident After Action Reviews (MATOC) Overview The MATOC Program coordinates the incident response activities in Northern Virginia, the District of Columbia, and Maryland—otherwise known as the National Capital Region. One of their essential duties is to aid the region in conducting incident after action reviews (AARs) for major incidents that affect multiple jurisdictions. They are charged with conducting the analysis and developing AAR performance summary reporting templates that can be used for communicating with responders, elected officials, and the public. C H A P T E R 4 Case Examples

22 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Main Use Secondary Use Title Agency Page # Performance Management and Safety Conducting Incident After Action Reviews Metropolitan Area Transportation Operations Coordination (MATOC) 34 Incident Notification and Bridge Closure Mitigation Pennsylvania DOT 39 Supporting Operational Decisions Using Probe Data Massachusetts DOT 41 Operations Traffic Operations Assessment Systems Tool (TOAST) Ohio DOT 43 Operations Winter Weather Maintenance Ohio DOT 48 Public Information Before and After Studies / Project Assessment New Jersey DOT 50 Operations Quantifying the Causes of Congestion U.S. DOT 52 Evaluating Effects of Pavement Restriping Projects Massachusetts DOT 55 Project Selection Virginia DOT 58 Congestion Management Process Richmond Regional TPO 60 Traveler Information Operations Holiday Travel Forecasting Georgia DOT 62 Safety Operations, Congestion Management Barrier Height Congestion Analysis Louisiana DOTD 66 Speed Data Use Cases Performance Management Operations Planning Table 5. Speed data use cases by category of use. Main Use Secondary Use Title Agency Page # Bike Evaluating Economic Impact of Bicycle Facilities Hampton Roads Transportation Planning Organization 68 Transit Transit Planning Massachusetts DOT 71 Freight Hampton Roads Port Truck Movement Hampton Roads Transportation Planning Organization 74 Freight Freight generator facility analysis Rhode Island Department of Administration 77 Work Zone Monitoring, Planning Work Zone Detour and Impact Analysis Maryland DOT State Highway Administration 80 Signal Operations, Planning Intersection Performance Analysis: Percent arrivals and green and turning movements City of Austin, TX 82 Pandemic Response Planning Analyzing people movement before and during the COVID-10 pandemic U.S. DOT, Maryland State Highway Administration and the Maryland Transportation Institute 86 Trips and People Movement Data Use Cases Planning Operations Table 6. Trips and people movement data use case by category of use.

Case Examples 23 Data Source Description The following data sets were used by MATOC to perform this AAR: 1. Third-party probe-based speed data. 2. Incident information including: location information, lane closures, responder information, vehicles involved, operators’ communications logs, and dynamic message sign messaging. 3. Closed-circuit television screen captures. Tools and Analysis Description MATOC used several tools with the RITIS Platform to analyze incident response and the resulting traffic impacts. These tools included the Incident Timeline viewer, Event Query Tool, and Probe Data Analytics. Below are descriptions of the tools that they use, and the types of outputs they provide. Incident Timeline Viewer The Timeline tool allows MATOC to see what actions were taken during the incident and the corresponding impacts to congestion. As shown in Figure 7, each panel in this tool shows inter- connections between different actions at any point of time. For example, communications in the top pane can be reviewed in conjunction with responses in the second pane. The responses visually indicate when each responder was notified (first diamond), the amount of time for them to arrive on the scene (dotted line), and the amount of time they spent on the scene (solid line). Similarly, the third pane shows lane closures over time and fourth pane shows when different messages were posted to relevant DMS. Finally, the bottom pane uses color to show probe vehicle data indicating level of congestion as well as any other events occurring in the area and upstream and downstream from the incident. The congestion caused by the closure of the majority of the lanes resulted in 10-mile queues upstream of the incident. Trend Map The Trend Map tool shows congestion impacts resulting from the incident as compared to a baseline day. The tool animates congestion, which is computed as a percentage of the free flow speed, and visualizes multiple days side by side to point out abnormal delays. In this example, MATOC evaluated congestion patterns in the immediate area of the Woodrow Wilson Bridge during the day of the incident compared to travel times on the same days of the week several weeks prior. At 8 p.m. on June 20, the congestion was still extending upstream from the incident on the main corridor as well as several surrounding corridors, compared to no congestion at the same time the previous weeks, as shown in Figure 8. User Delay Cost The User Delay Cost (UDC) tool shows the financial impact of delays caused by this incident, as shown in Figure 9. MATOC was able to analyze the entire month of June 2018 to compare user delay cost patterns on one of the detours. User Delay Cost on this alternate route was nearly $2M higher than it would normally have been. Travel Time Performance Charts MATOC uses performance charts to compare the travel times on roads surrounding the incident to understand the impacts the day of the incident compared to normal congestion. Figure 10 shows the travel time index was nearly eight times higher on the day of the June 2018 Woodrow Wilson Bridge incident compared to normal on one of the detour routes used to avoid the incident.

Figure 7. Example of an incident timeline from a June 2018 major incident on the Woodrow Wilson Bridge at the borders of the District of Columbia, Maryland, and Virginia. The bottom pane is a congestion heat map of queue build-ups during the course of the incident (Source: MATOC).

Case Examples 25 Figure 8. Animated congestion maps allow MATOC to understand roadway performance on the day of the incident compared to typical traffic (Source: MATOC). Figure 9. UDC analysis using probe-based speed data and volumes (Source: MATOC).

26 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Results and Outputs The outputs of the MATOC AARs include slide decks, recorded presentations, and brochures like the one seen in Figure 11, which is a four-page brochure that summarizes the incident, the effects on traffic, and recommendations. The AARs produced by MATOC are frequently requested by public officials and DOT execu- tives. To help their staff cope with the requests, they have produced tutorials on how to quickly Figure 10. Graph of travel time index on the George Washington Memorial Parkway. The red line is the travel time index for the day of the incident. The green and blue lines are from the prior weeks (Source: MATOC). Figure 11. Four-page AAR flier produced by MATOC.

Case Examples 27 Fractured steel truss under bridge Figure 12. Location of the fractured steel truss under the Delaware River Bridge in Pennsylvania (Source: Pennsylvania Turnpike Commission). put together slide decks and multi-page brochures using the RITIS platform. These types of after action reviews occur frequently for many different types of incidents including major weather events and special events. A recorded presentation by a MATOC staffer of the June 2018 major incident can be viewed at https://vimeo.com/364158242. This presentation shows more closely the types of data outputs produced through the analysis as well as some of the specific recommendations as a result of the AAR. Incident Notification and Bridge Closure Mitigation (PennDOT) Overview On January 20, 2017, the Pennsylvania Department of Transportation (PennDOT) identified a fractured steel truss under the bridge decking on the Pennsylvania side of the Delaware River Bridge, as shown in Figure 12. This discovery required full closure of the bridge to perform repairs. Traffic was detoured to the Pennsylvania Turnpike and the New Jersey Turnpike. To better manage this closure and potential impacts, PennDOT District 6-0 Regional Traffic Management Center (RTMC) used real-time data and tools to monitor regional traffic and dispatch service patrol vehicles to incidents as necessary. Data Source Description PennDOT District 6-0 RTMC used INRIX probe-based speed data and CCTV camera streams to monitor conditions and dispatch service patrol units. Tools and Analysis Description PennDOT used map-based visualizations with probe-based speed data, as shown in Figure 13, to evaluate congestion patterns before the closure and during the first week of the closure. PennDOT used probe data to build a customized dashboard to monitor travel time on detour routes and anticipated alternate routes of travel. Operators defined several travel time widgets on the dashboard to monitor travel times during morning and evening peak periods. This dashboard was saved every 30 minutes and screenshots were emailed to PennDOT and PA Turnpike staff and

28 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation consultants. Dashboard data were also saved for the period between January 23 and February 6, 2017, for the duration of the closure. In addition to evaluating the overall performance of the regional system, operators also used the travel time dashboard, as shown in Figure 14, to detect incidents. If travel time suddenly spiked, it indicated a sudden slowdown and potential incident. Results and Outputs PennDOT was able to quantify the effects of the PA Turnpike closure and confirm the Traffic Management Plan. Data were used to mitigate the effects of the closure, as well as manage alter- nate routes. For example, an increase in travel time on PA-413 initiated signal-timing changes Figure 13. Snapshot of congestion in the area of the bridge closure during the morning peak period (Source: Pennsylvania Turnpike Commission). TMC Operators used probe data to identify travel time impacts as a result of an incident. The probe data alerted them before traditional CAD sources. Figure 14. Screenshot of the speed data dashboard used to identify incidents (Source: Pennsylvania Turnpike Commission).

Case Examples 29 that were then monitoring using the same travel time dashboard. RTMC operators were able to identify incidents by observing sudden spikes in travel times. In addition to real-time monitoring, PennDOT was able to use data to support a multi- jurisdictional management plan and support future major events coordination. Supporting Operational Decisions Using Probe Data (MassDOT) Overview To provide additional capacity in peak hours, the Massachusetts Department of Transportation (MassDOT) has implemented a movable barrier system (also called “zipper lanes”) to allow for contraflow traffic on I-93 south of Boston. This system provides an additional lane of traffic in the northbound direction during the morning peak hours and an additional lane of traffic in south- bound direction during the evening peak hours to alleviate congestion. To enable contraflow, a special machine traverses the roadway and picks up the barrier, passes it through the conveyer belt, and sets it in the correct location to create another lane. As a result of the COVID-19 outbreak, there was an anticipated reduction in traffic on the roadways and MassDOT wanted to find out how that may affect their standard daily operations, while continuing to protect their DOT employees in the field. Data Source Description MassDOT primarily used INRIX probe data to observe historical patterns, recent patterns, and real-time flow metrics. They supplemented these data with traditional traffic counts where avail- able to account for the volume of traffic in the area. Tools and Analysis Description MassDOT engineers used RITIS Probe Data Analytics tools to graph travel times along the corridor and compare those travel times across multiple time periods. Results and Outputs MassDOT engineers observed morning peak travel time along I-93 northbound south of Boston on the Thursday after the governor’s stay-at-home advisory was executed and compared it to the average of Thursdays during the same period over the previous two years. They found that the average morning peak travel time was as much as 26 minutes shorter than during the previous two years, as shown in Figure 15. In addition to comparing raw travel times, the engineers also evaluated congestion patterns along the corridor over time using the Congestion Scan tool shown in Figure 16 that clearly showed significant reduction in congestion duration and length. As a result of this observation, MassDOT decided to suspend zipper lane operation, as it was unnecessary due to reduced demand in peak hours. Following the same logic, MassDOT intends to utilize probe data to evaluate travel times and congestion patterns during the pandemic against historical trends to guide their decisions regarding lane closures for scheduled and emergency roadwork to reduce costs and exposure for workers. Similarly, they intend to use these metrics and methods when businesses begin reopening to better assess demand and flexibility in operations.

30 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Figure 15. Average peak travel times on I-93 northbound compared between equivalent days in 2018, 2019, and 2020 (Source: MassDOT). 20182019March 26, 2020 Figure 16. Comparison of levels of congestion during a.m. peak on I-93 northbound in 2018, 2019, and 2020 (Source: MassDOT).

Case Examples 31 Traffic Operations Assessment Systems Tool: TOAST (Ohio DOT) Overview The Ohio Department of Transportation (ODOT) has implemented many ITS system upgrades over the years, but has done so in a largely ad-hoc fashion. At the same time, ODOT has been aggressively tracking numerous performance measures ranging from basic travel time perfor- mance reports to a patented snow and ice evaluator. The relative ease of visualizing and reporting traffic impacts, in addition to increased demand from motorists to solve congestion problems today elevated the need for ODOT to begin making data-driven decisions. ODOT’s Transportation Systems Management and Operations Program Plan recommended the development of a Traffic Operations Assessment Systems Tool (TOAST) to help the agency systematically assess operationally sensitive segments of roadway. The TOAST tool was to combine seemingly disparate sources of data to measure and rank traffic operations perfor- mance on each segment. Data Source Description ODOT staff prioritized the following data sources for integration into their tool: 1. Probe-based speed data: used to measure bottlenecks and travel time performance. – Bottlenecks � A potential bottleneck is detected when speeds on a segment drop to 65% of reference speeds and cause at least a 2-minute delay. – Travel time performance � Percentage of time motorists can travel at or near (90%) of the reference speed. 2. Crash report data: used to measure various safety measures and traffic incident management performance. – Safety performance � A route’s potential for safety improvement by density based on its peer group. – Incident clearance � The time from the report of an incident (crash) until the entire scene is cleared. – Secondary crashes � Percentage of crashes that occurred as a result of a previous incident. 3. Volume and count data. – Volume per lane – Freight corridors � Percentage of trucks traveling each segment. Tools and Analysis Description ODOT developed TOAST in-house, initially using Excel formulas and spreadsheets with custom- izable weighting and categories which can adjust to each District’s/program’s needs. The two probe- based speed data measures were given the heaviest weighting as shown in Table 7. However, TOAST has now been upgraded to run directly off of ODOT’s TSMO Data Warehouse where calculations are now automated, requiring far less interaction and manual data entry. Each performance measure is calculated on 1 year’s worth of data and combined to create a single road “segment” score. A lower TOAST segment score represents worse performance. More than 4,000 TOAST segments exist within the tool—providing a fine granularity of performance across the system and allowing operators to better understand how performance on one segment may be affecting adjacent segments or entire corridors. The segments are further broken down by

Min Max 0 15,000 10 15,000 18,500 9 18,500 22,500 8 22,500 30,000 7 30,000 35,500 6 35,500 50,000 5 50,000 65,000 4 65,000 86,500 3 86,500 122,000 2 122,000 170,000 1 170,000 0 0% 66% 0 66% 72% 1 72% 76% 2 76% 80% 3 80% 82% 4 82% 85% 5 85% 88% 6 88% 91% 7 91% 94% 8 94% 98% 9 98% 100% 10 0 0 10 0 0.6 9 0.6 1 8 1 2.0 7 2.0 3.5 6 3.5 6 5 6 9.0 4 9.0 17.0 3 17.0 34 2 34 85 1 85 0 0 2,200 10 2,200 2,700 9 2,700 3,300 8 3,300 3,900 7 3,900 4,400 6 4,400 5,400 5 5,400 6,600 4 6,600 9,700 3 9,700 15,300 2 15,300 30,500 1 30,500 0 0% 3% 10 3% 6% 9 6% 8% 8 8% 10% 7 10% 12% 6 12% 14% 5 14% 16% 4 16% 18% 3 18% 20% 2 20% 22% 1 22% 0 0 45 10 45 50 9 50 60 8 60 65 7 65 75 6 75 80 5 80 85 4 85 90 3 90 100 2 100 115 1 115 0 0% 5% 10 5% 6% 9 6% 7% 8 7% 8% 7 8% 9% 6 9% 11% 5 11% 13% 4 13% 15% 3 15% 17% 2 17% 23% 1 23% 0 100% 10 100 20 Multiplier Max Score 25.0% Bottlenecks SUM of Top 5 Impact Factors Impact Factor = Avg duration (minutes) x Avg max length x Number occurrences 0-170,000+ 2.5 25 Weighting % Criteria Calculated by Value Range Range Normalized Value 20.0% Travel Time Performance Score (%) Real Travel Time/Target Travel Time is ≥ 90% of Reference Speed 0-100% 2.0 15 15.0% Safety Performance Potential for Safety Improvement by density (PSI_Density) Crashes per year per mile 0-85+ 1.5 15 15.0% Volume per Lane Number Vehicles 0 - 30,500+ 1.5 7.5 10.0% Freight Corridors Percent Trucks (%) 0-22%+ 1.0 10 7.5% Incident Clearance Average Duration (minutes) 0-110+ 0.75 7.5% Secondary Crashes Ratio of Secondary Incidents to Total (%) 0-33%+ 0.75 7.5 Use Criteria? TRUE Data Date Ranges 2019 Fiscal Year (7/1/18 - 6/30/19) 2019 Fiscal Year (7/1/18 - 6/30/19) 2018 Calendar Year (1/1/18 - 12/31/18) 2018 Calendar Year (1/1/18 - 12/31/18) 2018 Calendar Year (1/1/18 - 12/31/18) 2018 Calendar Year (1/1/18 - 12/31/18) 2018 Calendar Year (1/1/18 - 12/31/18) Table 7. ODOT’s weighting of each performance measure (Source: ODOT).

Case Examples 33 functional class, urban/rural classifications, and intersection/interchange center points. This finer granularity and classification system helps to prevent operational performance from becoming washed out in the data and key problem areas being overlooked. Results and Outputs TOAST ranks the performance of every segment and corridor in a list, as shown in Figure 17. To communicate to different stakeholders and meet different user needs, different types of ranked lists are provided, including the • Statewide Top 50 list, • Urban Freeway list, • Urban Arterial list, • Rural Freeway list, and • Rural Arterial list. These lists are further broken down by each of ODOT’s 12 Districts. Static maps are generated for each district and county to convey performance graphically, as shown in the example in Figure 18, and a publicly accessible interactive statewide map is also published, as depicted in Figure 19. Based on the TOAST data, ODOT District TSMO Coordi- nators and consultants developed TSMO studies. The first TOAST list and subsequent TSMO Studies produced three pilot projects. The second annual TOAST list, released in 2019, generated funding for 10 TSMO projects. ODOT staff members believe that the implementation and use of TOAST has had a signifi- cant impact on their TSMO Program. The development and release of TOAST were directly responsible for ODOT receiving dedicated, programmatic-level TSMO funding. TOAST further supports the agency’s planning practices to increase the focus on lower cost and quicker Figure 17. Screenshot of ODOT’s Statewide Top 50 list (Source: ODOT).

34 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Figure 18. PDF maps of segment performance are provided to each district (Source: ODOT). implementation countermeasures. When evaluating upcoming construction projects, dis- tricts leverage TOAST to understand if a TSMO implementation should be included. When evaluating similarly-scored, competing projects, the use of TOAST can often mean one project wins over another. Other programs have benefited from TOAST including ODOT’s Traffic Incident Manage- ment (TIM) program. The TOAST tool was used to rank roadway segments based on secondary crashes and incident clearance times. This new ranked list is used to help inform TIM practices in the state. ODOT shares TOAST with partners such as cities, MPOs, and others. Some partners are modifying the TOAST concept to build similar programs and tools for their own use. More information on TOAST can be found at https://www.transportation.ohio.gov/wps/ portal/gov/odot/programs/tsmo/resources/toast.

Case Examples 35 Winter Weather Maintenance Performance Management (Ohio DOT) Overview ODOT was challenged with enabling travelers to continue to utilize roadways even after major winter weather storms. ODOT developed a specific activity and outcome-based performance measure called “recovery time,” which allowed them to continually monitor their operational strategies during winter weather events. The measures incentivize recovery efforts and commu- nicate performance to management and the public. Data Source Description ODOT invested in probe-based speed data from a third-party provider after abandoning traditional stationary speed sensors due to issues with spatial resolution and costs. The data are updated by the provider every minute, and includes speed, travel times, and confidence scores on all roadway segments. Figure 19. Screenshot of the ODOT’s Transportation Information Mapping System accessible at https://gis.dot. state.oh.us/tims/map.

36 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation ODOT also collects roadway conditions via Road Weather Information Systems (RWIS) data to identify when a storm begins or ends. Tools and Analysis Description The ODOT dashboard was created in-house by agency staff. Using information from the stations, they can determine when a snow or ice event begins and/or ends. A snow or ice event begins when 40% of a County’s RWIS stations detect either snow or freezing rain, paired with the following criteria: • The air temperature or pavement temperature is below 34°, and • The speed drops more than 10 mph below its expected value on at least 25% of designated routes (minimum two) within the county. A snow and ice event is considered complete when the following are true: • At least 60% of the county’s RWIS stations are reporting “None” or “Rain” as the precipita- tion type, • The wind speed detected from RWIS stations drops below 15 mph (to account for drifting snow), and • A new snow and ice event does not begin within another 2 hours. When the above conditions are met, ODOT begins tracking performance conditions and the time it takes to reach certain performance goals. The time from the end of the event until the speeds recover is called the “recovery period.” Each county in the state has a recovery goal on each designated route. The goal is 2 hours for routes designated as a Priority 1, and 4 hours for Priority 2 routes. The recovery period officially ends once the speeds recover to within 10 mph of their expected values for at least an hour. Any route that has not recovered within the goal will be reported. Results and Outputs The visibility of their dashboard by management and analysts within ODOT meant they have been able to implement operational strategies that are improving performance and customer satisfaction. ODOT can focus on areas with the slowest recovery period and/or districts that struggled to recover all of their routes to deploy additional roadway treatments for future storms. ODOT also leveraged existing RWIS data to evaluate roadway conditions (like salinity content and surface temperature), and a combination of CCTV and probe-based speed information to further identify potential sources of slow recovery time and dispatch service patrols to assist stranded motorists, or direct travelers to alternate routes through traveler information resources and media campaigns. ODOT’s interactive dashboard also allows for drill-down capability, as shown in Figure 20, which enables users to look at specific districts and specific event performance. This type of drill- down capability made the measures much more insightful as it began to answer the question of why certain districts and routes were performing worse than others. Other agencies are also adopting similar measures with different names or modified defini- tions. The Michigan DOT, for example, has implemented a “regain time” measure that is similar to Ohio’s Recovery Period measure. FHWA’s Best Practices for Road Weather Management case study discusses the Michigan implementation and can be found at https://ops.fhwa.dot.gov/ publications/fhwahop12046/rwm16_michigan1.htm.

Case Examples 37 Project Assessment and Before-and-After Studies (NJDOT) Overview The New Jersey Department of Transportation (NJDOT) wanted to increase transparency regarding the selection and impact of projects implemented throughout the state. To achieve this transparency, NJDOT needed to be able to analyze individual projects, quantify impacts, and communicate those impacts in a meaningful, reproducible, and understandable way. Prior to 2015, this was difficult to do due to the high cost of conducting analysis, the lack of necessary before-and-after data, and the lack of an ability to distill complex project and mobility data into a digestible format for the public and decision makers. Data Source Description To develop this capability, NJDOT used probe-based speed data for speed and travel time information and derived metrics, as well as NJDOT collected incident, event, and construction data. Volume data from agency sensors and studies were also used to help derive Level of Service (LOS) numbers. Tools and Analysis Description NJDOT used the Probe Data Analytics Suite to analyze congestion data around a recently completed project. They compared conditions prior to the project with those following Figure 20. Drilling down in the dashboard lets the user see how each event was managed, whether it hit the 2-hour recover period, or whether the recovery goal was missed. The one shown in the figure did not meet the 2-hour recovery timeline. Clicking on a specific location draws a diagram of the road (Source: ODOT).

38 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation completion, evaluating Vehicle Hours of Delay, Travel Times, Reliability, Speed Breakdowns (the percentage of time speeds was above or below various thresholds), and LOS. Results and Outputs Using the probe data and related tools, NJDOT has been able to realize significant cost savings through the system’s ease of use as well as through the ability to do the analysis quickly in-house as opposed to relying on consultants. NJDOT conservatively estimated that they would have normally had to spend 182 person-hours for in-house staff using traditional field investigation methods and developing summaries. If the agency were to conduct 25 similar project assess- ments or mobility reporting summaries a year, they estimate a total of $340K and 4,475 person- hours could be saved. If they were to compare these costs to farming out the work to consultants, they would save even more—close to $1M. In addition to cost savings, the agency saw an increase in the amount and quality of informa- tion that could be produced. Agency staff generated Project Assessment Summary brochures and pamphlets, as shown in the examples in Figure 21 and Figure 22, which provided plain language summaries of different projects and their costs and benefits. Figure 21. Graphic from the back of a Project Assessment Summary pamphlet comparing before-and-after conditions I-80 at Squirrelwood Road (Source: NJDOT).

Case Examples 39 Quantifying the Causes of Congestion (U.S. DOT) Overview The Bureau of Transportation Statistics (BTS), U.S. Department of Transportation, is working to update the 2004 Causes of Congestion Study (https://ops.fhwa.dot.gov/congestion_report_04/ index.htm) to leverage real-world data, updated methodologies, and better classifications. BTS is working with The Eastern Transportation Coalition (TETC), formerly the I-95 Corridor Coalition, to recreate a national-level pie chart, while also producing pie charts for all 50 states and Washington, D.C. The end product produced an interactive online tool for users to discover the national- and state-level causes of congestion, as well as the monthly congestion trends from each congestion category. An initial version of the tool is available now, with updated stats being deployed in August 2021. Data Source Description The BTS and TETC use data sources that cover the NHS. Table 8 summarizes the data sources used to identify congestion and lists the causes of congestion categories. Tools and Analysis Description BTS contracted with TETC to develop and implement the methodology seen in Figure 23. First, congestion is detected using probe-based speed data from a third-party provider. Statistical analysis determines if the congestion is recurring or non-recurring. If the congestion is non- recurring, then the other data sources identified in Table 8 are spatially and temporally fused to determine if some form of event (e.g., weather, work zone, or collision) may be to blame. If more than one potential cause is identified, then that is noted. Lastly, the amount of delay is quantified using volume data. Figure 22. Inside pages of the NJDOT Project Assessment Summary brochure showing results for I-80 at Squirrelwood Road.

40 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Data Item Data Source Data Size Congestion/Disruption 1-minute probe data 370K Highway segments with data for each minute Recurrent Congestion 1-minute probe data Incidents Waze 78M Waze Incident events Weather NOAA radar and Waze 5.6M Waze weather events and 2-minute radar readings for each 370K highway segment Work Zones Waze 8M Waze work zones Holiday Travel Holiday Calendar (including travel days before/after holiday) 46 holiday travel days Signals OSM Traffic Signal Database 332K traffic signals(each intersection approach was associated with a signal) Multiple Causes Combination of above a Unclassified Disruption 1-minute probe data a aThere is no easy way to quantify the size given how variable the data are from region to region and month to month. a Table 8. Data summary. Figure 23. Transportation Disruption and Disaster Statistics analytical framework (Source: TETC).

Case Examples 41 Results and Output The study team is still computing data for the entire country. The team expects to complete the computation in August 2021. The results of the analysis are provided to BTS and the public in CSV and Excel files. Early results for the states of Colorado and Maryland, shown in Figure 24 and Figure 25, are already posted online through an interactive website. A stakeholder group made up of DOT leadership from across the country is evaluating the results and providing guidance to BTS on next steps and desired enhancements. Figure 24. Preliminary results for Maryland (Source: TETC). Figure 25. Preliminary results for Colorado (Source: TETC).

42 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Evaluating Effects of Pavement Restriping Projects (MassDOT) Overview Many agencies struggle with the problem of recurring congestion in large metropolitan areas due to a lack of sufficient capacity to handle the demand, especially in peak hours. Agencies generally implement different operational strategies to better manage congestion. MassDOT implemented several pavement-restriping projects in the state to better manage the flow of traffic in congested areas and in large rotaries. Case 1 – Storrow Drive Tunnel In the first case, MassDOT was looking at Storrow Drive westbound in the area between Leverett Circle and Charles Circle in Boston, where traffic coming out of the Storrow Drive tunnel from I-93, the Tobin Bridge and the Sumner Tunnel joins traffic from the Leverett Circle on-ramp for a short span before splitting again for Charles Circle to the left and continuation of Storrow Drive to the right. Due to major inflow of traffic, the required lane changes, merging, and weaving in a short span, navigating the area was difficult especially for those unfamiliar with the area. Case 2 – Middleborough Rotary While rotaries generally accommodate multiple lanes of traffic, they often do not have pavement markings. MassDOT was looking at the case of the Middleborough rotary, which handles traffic to and from several major arterial routes, often resulting in congestion in and around the rotary and queuing on approaches to the rotary. Data Source Description MassDOT used INRIX probe data to evaluate the effects of pavement restriping projects. In addition to measuring travel time changes, MassDOT also evaluated changes in the reliability of travel time calculated as a difference between the 5th and 95th percentiles of travel time. Tools and Analysis Description MassDOT engineers used RITIS Probe Data Analytics tools to evaluate facility performance before and after projects were executed. The most effective tool in the toolset was a box and whisker plot of travel times that shows key travel time metrics: • Median travel time. • First quartile (25th percentile) travel time. • Third quartile (75th percentile) travel time. • Fifth quartile (95th percentile) travel time. This plot allowed engineers to calculate change in travel time as well as change in the reliability of travel time. Results and Outputs Case 1 – Storrow Drive Tunnel In this case, MassDOT modified pavement markings to eliminate the left lane trap (exiting to Charles Circle), which reduced the number of necessary lane changes and driver confusion, especially for those unfamiliar with the area. In effect, these pavement markings allowed many

Case Examples 43 drivers to remain in their lanes and not get stuck in the left lane needing to make last-minute lane changes. Prioritizing certain lane movements resulted in smoother flow of traffic. As shown in Figure 26 and Figure 27, while the travel time remained largely unchanged, with a slight increase (less than 30 seconds) in some approaches in the 6 a.m. peak hour, the travel time reliability improved significantly. Travel time was reduced significantly in the range between the 5th and 95th percentiles after restriping. While reduction of travel time is a good measure of improvement in throughput, improve- ments in travel time reliability often have an even more significant impact on traveler experience. Consistent travel time allows travelers to better use their time as they can better predict expected travel times. For freight operators, travel time reliability allows them to remain competitive because they can be more efficient in their pickups and deliveries. Case 2 – Middleborough Rotary In this case, MassDOT placed lane markings approaching and within the rotary that provided information regarding which lanes to be in for different destinations. This allowed drivers to remain in the proper lanes and eliminated lane changes in the rotary. MassDOT engineers evaluated changes in the travel time and travel time reliability before and after the placement of the pavement markings. In this case, both travel time and reliability improved on the approaches and through the rotary, especially in the hours between 12 p.m. and 6 p.m. Some travel time reductions were as much as 4 minutes. These reduced travel times indicate reduction in congestion and queuing, and reliability improvements provide better traveler experience and efficiency. Figure 26. Comparison of travel time on the affected segment of road for the same time period in 2018 and 2019 (Source: MassDOT).

44 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Project Selection (Virginia DOT) Overview The Virginia Department of Transportation (VDOT) is faced with the challenge of selecting and prioritizing projects given that there is never enough funding available to address all the needs of the state. Virginia’s Commonwealth Transportation Board (CTB) found that prior to 2014, the project selection process in the state was politically driven, not sufficiently transparent. This combination created uncertainties for local communities and businesses. That same year, the state of Virginia enacted SMART SCALE (VA House Bill 2 Legislation) requiring the CTB to develop and implement a quantifiable and transparent transportation funding process. Data Source Description To support effective decisions regarding funding, the SMART SCALE process relies on several different sources of supporting data, including the following: • Safety data, • Congestion mitigation data, • Accessibility data, • Environmental quality measures data, • Economic development measures data, and • Land use coordination data. Figure 27. Comparison of travel time on the affected segment of road for the same time period in 2018 and 2019 (Source: MassDOT).

Case Examples 45 The following data support the process for congestion metrics: • Vehicle miles traveled (VMT) from VDOT’s Traffic Monitoring System. VMT is a product of hourly and directional distribution of traffic, annualized traffic volumes, and length of roadway segment. • Speed data from INRIX, a vendor of real-time speed data gathered from mobile devices, in-vehicle GPS, and road sensors. The speed data were accessed via RITIS at https://ritis.org, through a VDOT subscription service. • Vehicle occupancy data from the 2018 FHWA National Household Travel Survey (NHTS) and the Virginia supplement (additional surveys conducted) to the NHTS. • Speed Limit Data from VDOT. Tools and Analysis Description The SMART SCALE process (http://vasmartscale.org/) consists of a methodology that includes a well-defined set of inputs, calculations, factor weighting, and scoring to produce a list of scored projects for the CTB to prioritize. The multi-step process of project assessment and prioritization is executed by multiple groups including a technical evaluation team and an external stakeholder review team. As part of this process, the teams use a mix of in-house databases and commercial-off-the-shelf tools, as well as hosted analysis systems that provide metric calculation and visualization capabilities. One of the many steps of the process includes measuring congestion for Interstates and limited-access facilities, and noting which sections of roadways are excessively congested, and by how much, as shown in Figure 28. Figure 28. Percentage of person miles traveled in excessively congested conditions for interstates and select limited-access facilities in the Richmond district (Source: https://vtrans.org/resources/Methodology_VTrans_ Midterm_Report_Richmond.pdf).

46 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation A description of the method used for the Richmond Construction District for 2019 can be found at https://vtrans.org/resources/Methodology_VTrans_Midterm_Report_Richmond.pdf. Results and Outputs Using the SMART SCALE data-driven process, the Commonwealth has been successful in making the process transparent and clear for project sponsors and the public. In addition to introducing accountability, the SMART SCALE process also identified improvements, such as emerging data sources that could be used to make even better project prioritization decisions. Congestion Management Process (Richmond Regional TPO) Overview Every urban area of 200,000 in population must implement a Congestion Management Process. The Richmond Regional Transportation Planning Organization (RRTPO) is responsible for devel- oping a Congestion Management Process (CMP) for the Richmond, VA region. The CMP tracks transportation system performance measures, outlines strategies to manage demand, and works to address multimodal reliability and sustainability issues. Implementing a CMP can be complex, and requires a significant amount of data analysis. The Richmond MPO decided to leverage probe-based speed data heavily for their more recent CMPs. Data Source Description The RRTPO uses a mix of probe-based speed data sources. They have access to 1-minute speed data from INRIX because VDOT purchases data on behalf of all state partners. The MPO also has access to the freely available NPMRDS probe-based speed data that are provided at 5-minute intervals. The combination of these two data sets covers all of the roads included in the CMP. The MPO is also working with VDOT to include crash data and other event data in the next CMP. Tools and Analysis Description The MPO leveraged third-party tools from the RITIS Platform to compute, download, and further analyze performance measures including but not limited to: • Travel Time Index. • Level of Travel Time Reliability (LOTTR). • Truck Travel Time Reliability (TTTR). • Bottlenecks (recurring and non-recurring). Recurring congestion is analyzed by monitoring the travel-time indices and through bottle- neck ranking reports embedded in the RITIS Platform. Non-recurring congestion is analyzed by monitoring the federally mandated performance measures (LOTTR, percent person-miles reliable, and TTTR) that are provided to the MPO by VDOT. They also leveraged Excel, ArcGIS, and other tools to refine their analysis and present their findings online. Results and Outputs Since leveraging the probe-based speed data in 2014, the agency has seen the quality of its analysis and the speed at which it can update the CMP increase. This has afforded the agency

Case Examples 47 more time to increase the quality of the presentation of the materials and free up time to identify and assess strategies to address congestion and mobility issues in the region. RRTPO publishes updated CMP analysis via ESRI’s Story-map Product, as shown in the example in Figure 29. The 2019 CMP can be viewed at https://planrva.maps.arcgis.com/apps/ Cascade/index.html?appid=b2d655a0bd774a6c84dd8f1672118f08. Holiday Travel Forecasting (Georgia DOT) Overview The Georgia Department of Transportation (GDOT) wanted to provide better guidance to the traveling public for what to expect during the week of major holidays like Thanksgiving. The agency hoped that the provision of better information would help travelers avoid peak periods of congestion and reduce overall delays. Data Source Description GDOT purchases real-time probe-based speed data from a third-party data provider. These data are archived for several years in the RITIS Platform at 1-minute intervals. The data cover Figure 29. Screenshot of the 2019 RRTPO CMP update (Source: https://planrva.maps.arcgis.com/apps/Cascade/ index.html?appid=b2d655a0bd774a6c84dd8f1672118f08).

48 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation the bulk of the roads in the state and are also used for other transportation operations purposes, including traveler information and situational awareness. Tools and Analysis Description Using in-house staff, GDOT looked at travel patterns during prior year’s Thanksgivings and analyzed the travel statistics for different parts of Atlanta along with the region as a whole. GDOT staff used the RITIS Probe Data Analytics Platform to conduct the analysis and build some of the end visualizations. GDOT modified existing Holiday Travel Forecast templates from the RITIS website to give them GDOT-specific branding and style. Results and Outputs The end result of GDOT’s analysis was a series of infographics that resembled weather reports, but for traffic. These graphics, seen in Figures 30–32, were distributed via the agency’s website, press releases, and even Twitter. Many media outlets in the region ran the story and interviewed GDOT staff, as can be seen at https://www.wsbtv.com/news/local/the-best-and-worst times-to-travel-in-atlanta-this-thanksgiving/876096230/. Figure 30. GDOT tweet of an animated holiday traffic map (Source: GDOT / Twitter).

Case Examples 49 Figure 31. Region-specific travel forecasts for the 2019 Thanksgiving Holiday week (Source: GDOT).

50 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Figure 32. Example of a more stylized single-region travel forecast (Source: GDOT). Barrier Height Congestion Analysis (LADOTD) Overview The Louisiana Department of Transportation and Development (LADOTD) has been building tall median separation barriers (54″ tall) as the standard for many years; however, there is a significant construction cost to these taller barriers compared to the shorter (32″ tall) barriers. LADOTD is exploring the idea of switching to the shorter barriers in some cases throughout the state, but there are concerns about the impacts on both safety and congestion. Engineers wanted to know if the larger barriers would have an impact on gawker-related congestion, which is congestion resulting from drivers on the opposite side of the roadway slowing down to look at incidents and activity on the other side of the road. Data Source Description The following data sets were used by the Louisiana Transportation Research Center (LTRC) for this technical assistance task: 1. NPMRDS probe-based speed and travel time data. The NPMRDS was the only wide- area speed data source available to LADOTD at the time of this analysis. Even with the 5-minute granularity and occasional coverage gaps, the data were still adequate to show congestion issues. 2. The locations of different barrier heights, including those deemed less effective despite their actual heights. For example, as shown in Figure 33, vertical curves, grade differences, and glare screens may all have high barriers, but they may not be as effective. 3. The locations and timestamps of incidents with the potential to result in gawking (including crashes, overturned vehicles, vehicle fires, spills, and roadway debris). Tools and Analysis Description LTRC temporally and spatially conflated LADOTD’s incident data with the NPMRDS data downloaded from npmrds.ritis.org for the Interstate locations in the Baton Rouge area where existing median barriers were present. LTRC constructed a Tableau dashboard that allowed them to screen the NPMRDS data for delays that met all six of the following conditions: (1) an incident was present; (2) there was a speed drop in the opposite direction; (3) there was no recurring congestion; (4) there was no congestion immediately downstream; (5) the incident

Case Examples 51 Figure 33. Examples of Lower Effective Barrier Heights (Source: Kirk Zeringue, LTRC). was not weather-related; and (6) the incident was not work-zone or special event related. If all six conditions were met, then the associated incident was deemed the only detectable source of delay in the opposite direction. Each detected location was then evaluated based on the field conditions to determine if a condition existed resulting in a lower effective median barrier height for those incidents in the area of tall barriers (54″). Results and Outputs For the Baton Rouge case study, LADOTD determined that at least 26 of the 28 opposite- direction rubbernecking incidents that occurred on I-10 and I-12 during the study dates occurred in locations with either a 32″ barrier or a 54″ barrier with lower effective median barrier heights, as shown in Figure 34. This information is now being used to complement the case being made for the continued use of tall barriers in Louisiana. Figure 34. Map of barrier heights along I-10 and I-12 with event locations (Source: Kirk Zeringue, LTRC).

52 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Figure 35. Strava heat map for Historic Triangle (Source: https://www.strava.com/heatmap#14.00/ -76.26597/36.86823/hot/all). Evaluating the Economic Impact of Bike Facilities (Hampton Roads, VA) Overview The Hampton Roads Transportation Planning Organization (HRTPO) wanted to study the potential of bike facilities to improve the lives of residents and promote social development. The first phase of their study included an economic impact analysis of the bike facilities in Hampton Roads, beginning with the identification of key facilities, usage patterns, and models to use to evaluate the economic impact of those facilities. Data Source Description HRTPO also used GPS-based data from two providers to better understand usage patterns. As depicted in Figure 35, one set of data was obtained from the Strava app, which collects GPS location information primarily for runners, walkers, and bicyclists. The other GPS-based data set was from StreetLight, which aggregates various LBS data from multiple sources. HRTPO used StreetLight to perform “Visitor Home and Work Analysis.” Tools and Analysis Description HRTPO also used StreetLight analysis tools for Visitor Home and Work Analysis. Results and Outputs StreetLight data and analysis tools allowed HRTPO to analyze the home and work locations of visitors to a zone. They analyzed 12 months of data and around 8,500 unique devices separated into temporal buckets (average day, average weekday, average weekend, and parts of the day such as morning, mid-day, afternoon, evening, retail hours, and all day). As a result, HRTPO found that while the majority of the users of the Virginia Capital Trail come from Virginia (69%), one-third of all users are visitors from other states, with the top three visiting states being North Carolina (4%), Maryland (3%), and Pennsylvania (3%) as shown in Figure 36 and Figure 37.

Case Examples 53 Figure 36. Home states of Virginia Capital Trail users (Source: HRTPO analysis of StreetLight data, 2017). Figure 37. Trail users’ home locations in the Eastern United States (Source: HRTPO analysis of StreetLight data, 2017).

54 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation In addition to state-level statistics, HRTPO was able to extract additional information about non-local users from Virginia. The next phase of this study will use the findings from the first phase and perform additional surveys of dollars spent by visitors using the Virginia Capital Trail to evaluate the economic impact to the region. Additional information on the study can be found at https://www.hrtpo.org/uploads/docs/ Economic%20Impact%20of%20Bicycle%20Facilities%20in%20Hampton%20Roads-%20 Phase%20One.pdf. Transit Planning (MassDOT) Overview The Massachusetts Department of Transportation’s Office of Transportation Planning was evaluating the effectiveness of the current transit service and potential opportunities for service improvements and expansion. The goal of the study was to observe travel patterns between the city of Boston and metro regions, especially in peak hours. The study demonstrated that there is both sufficient demand for expanded transit service as well as opportunities to greatly improve service by reallocating existing general travel lanes to bus lanes. Data Source Description As part of this study, MassDOT used INRIX Trips data, in combination with LBS data, as well as some of the more traditional data sets such as traffic counts, transit schedule and performance data, Automated Fare Collection data, and Origin-Destination-Transfer data. INRIX Trips data consist of individual trips defined by origins and destinations and a set of segments between those origins and destinations. In addition to the specific path traversed, each trip also has additional information such as trip departure and arrival times and overall travel time. Tools and Analysis Description MassDOT used the RITIS Trip Analytics tool to evaluate trip patterns. This tool consists of three modules: (1) O-D Matrix, (2) Segment Analysis, and (3) Route Analysis, each with a different set of capabilities. As part of this transit study, MassDOT used Segment Analysis to quantify the number of trips between different Traffic Analysis Zones (TAZs) using segments of the Route 1A corridor. After selecting a single segment or set of segments, time period, and geographic resolution, MassDOT planners were able to visualize primary origins and destina- tions for trips traversing that set of segments and observe specific patterns. MassDOT planners also utilized the O-D Matrix module to determine the number of origins per municipality for a specific destination. Results and Outputs As a result of this type of analysis, MassDOT planners identified that in morning peak periods (6 a.m.–8 a.m.) on weekdays, between 11% and 14% of trips going southbound on the Lynnway (Route 1A) between Commercial Street and the General Edwards Bridge end in the TAZ that includes the Wonderland garage, as shown in Figure 38.

Case Examples 55 Wonderland Garage Location Lynnway between Commercial St. and General Edwards Bridge Figure 38. Number of trips traversing Lynnway segment and ending in the area of the Wonderland garage (Source: MassDOT). When medium and heavy vehicles are removed from the analysis, between 17% and 22% of trips going southbound on this segment end in the TAZ that includes the Wonderland garage. The Wonderland garage is served by the terminal station of the MBTA Blue Line, indicating that many of the users of this roadway transfer from their personal automobiles to transit in order to complete their commute. Approximately 55% of trips in light vehicles using this corridor end in East Boston, Downtown Boston, South Boston, and Kendall Square, and approximately 81% of trips end in Revere, Boston, and Cambridge. This indicates that much of the travel in this corridor could also be completed using transit services, and there are opportunities to encourage mode shift by improving transit services. This analysis was also conducted for the segment of North Shore Road (Route 1A) south of the General Edwards Bridge, as shown in Figure 39. This segment was chosen in order to exclude traffic that diverts to the parallel road, Revere Beach Boulevard. The results show that travelers who remain on this corridor could benefit from improved transit service on this portion of the corridor in addition to improvements on the Lynnway. In addition to analyzing the destinations of users of the Route 1A corridor, MassDOT Planners analyzed origins of trips along the corridor. The results indicate that origins are concentrated in the municipalities adjacent to Lynn. Figure 40 shows that by improving access to the MBTA Blue Line from the Lynn Central Garage, it may be possible to encourage these drivers to park at the garage and use improved bus services to access the MBTA Blue Line. Finally, MassDOT planners used the O-D Matrix module to determine the origins by munici- pality of trips ending in the TAZ that includes the Wonderland garage, further demonstrating

56 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Wonderland Garage Location North Shore Rd. at Point of Pines Figure 39. Number of trips traversing North Shore Rd. segment and ending in the area of the Wonderland garage (Source: MassDOT). Figure 40. Counts of a.m. weekday peak trips originating in municipalities adjacent to Lynn (Source: MassDOT). the potential of serving trips that use Route 1A by improving bus transit between the Lynn Central Garage and the MBTA Blue Line. Hampton Roads Port Truck Movement (HRTPO) Overview Seaports are a significant source of economic activity that heavily depends on the quality of transportation infrastructure in the port-surrounding area. The Virginia Commonwealth has several ports along its coast that produce a significant truck movement as goods and materials leave and arrive at ports. As part of project selection in a constrained budget environment, HRTPO wanted to identify key highway gateways used by port trucks to use as input for HRTPO’s Project Prioritization Tool and major regional studies in the state.

Case Examples 57 Data Source Description HRTPO used two separate sources of data. They first used American Trucking Research Institute GPS data that are based on 1-minute GPS pings from participating companies’ fleets. HRTPO believed these data were biased towards specific companies that participate in the pro- gram and therefore generated odd results that were not representative of the overall trucking movement. As shown in Figure 41, HRTPO then used StreetLight data sourced from connected vehicles, smart phones, and GPS-based fleet management systems. These data included location informa- tion and travel patterns between locations. HRTPO believed this source to be more representa- tive than the ATRI data. Figure 41. Destinations of trucks from port-related distribution centers, July 2016–June 2017 (Source: HRTPO staff processing, via ESRI, of trips from Port Distribution Centers via StreetLight).

58 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Tools and Analysis Description HRTPO analyzed StreetLight data using StreetLight’s web-based analytical tool InSight. HRTPO benefited from VDOT’s existing contract with StreetLight that made the data and tool available to HRTPO staff to perform analysis in-house. Results and Outputs After the initial analysis of the StreetLight data that looked at trips between local container port facilities and regional highway gateways, HRTPO staff realized that trips were terminated if a vehicle was stopped for 5 minutes. To overcome this limitation, HRTPO staff looked at trips between port-related distribution centers and regional highway gateways which allowed them to extrapolate which trips originated at the port and ended outside of the port area. Using this information, HRTPO staff members were able to identify the primary highways utilized by port truck traffic for each port in the commonwealth. For example, as shown in Figure 42, port-related distribution centers on the peninsula mostly use I-64 (93%) to leave the region. Similarly, port-related distribution centers on the Southside primarily use US-460 (39%) and US-58 (34%) to leave the region. This result is somewhat surprising given that many destinations lie west and north of Richmond and would logically require use of I-64. However, congestion at harbor crossings (the Hampton Road Bridge Tunnel and the Monitor Merrimac Memorial Bridge Tunnel) may detract trucks from using that route. Figure 42. Usage of gateways by Southside port-related distribution centers (Source: HRTPO staff processing of StreetLight data, July 2016 through June 2017).

Case Examples 59 Freight Generator Facility Analysis (Rhode Island DOA) Overview The Rhode Island Department of Administration (DOA) wanted to be able to determine the location of the top freight generators and destination patterns within and outside of Rhode Island. Data Source Description DOA purchased Trips data from INRIX. Part of this purchase included the Origins and Destina- tions of each trip broken down by light-, medium-, and heavy-vehicle classifications. The Origins and Destinations each had latitude and longitude points associated with them. They also used their own TAZ shapefile layer data. Tools and Analysis Description The DOA downloaded the full matrix of origins and destinations by TAZ layers from INRIX. Using TransCAD, they separated the data into two fields—Origins by Destination and then Des- tinations by Origin. They joined this to a TAZ centroid layer, and then used QGIS to create a basic heat map as shown in Figure 43. Figure 43. Heavy truck origins in Rhode Island (Source: Rhode Island DOA).

60 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation After generating the initial heat maps, DOA staff then conducted further drill-down analysis to analyze top truck-freight origin community maps, and pinpoint other top freight locations at a more granular level. Results and Outputs After DOA staff overlaid business location data onto their heat maps, they were able to quickly correlate freight generators with specific business names and types. For example, Figure 44 shows the DOA freight heat map for the Cumberland region and the matching businesses. The data proved valuable to DOA, and their analysis helped to validate the data and show where many of the data were coming from. DOA plans to use the data and insights from this analysis to help refine their new Truck Model for their State Travel Demand Model. They will also use the results to aid in outreach and coor- dination with local planners in the communities with freight generators to discuss issues such as truck parking, truck idling, air quality, noise, safety, and freight planning. They also plan to use the analysis to determine with whom to conduct additional outreach in the private sector for certain freight-related issues. For example, the DOA did not realize how much the beverage industry was contributing to freight in the region until after they analyzed the data. Work Zone Impact Analysis (Maryland DOT) Overview The Maryland Department of Transportation State Highway Administration (MDOT SHA) wants to understand how (or if) travelers detour around active work zones, major incidents, or even tolling facilities. For work zones and incidents where detours are often suggested, MDOT wants to know if their guidance is being followed, or if traffic is utilizing unintended routes like neighborhood streets that are not designed to sustain increased volume and weight. Through the analysis of past behavior, MDOT intends to eventually: 1. Predict behavior of traffic in and around planned work zones or incidents and turn that information into an exportable piece of data that can be used by third-party applications to inform travelers of their options and potential restrictions, FedEx Facility Figure 44. DOA heat map and analysis identifying the heavy freight generator FedEx (Source: Rhode Island DOA).

Case Examples 61 2. Balance demand on arterials, 3. Develop more effect communication strategies for DMS, media, websites, and third-party traveler information providers, and 4. Use historical work-zone performance data combined with historical and real-time traffic information to better manage permitting and work-zone activation to reduce the overall conflict between work-zone closures and non-work-zone-related congestion. Data Source Description For this project, the MDOT SHA will leverage three primary data sources: 1. INRIX Trips data, which provide the origin of trips, the destination of trips, and the routes taken between the origin and the destination. These data will be used to understand route choice and detours. 2. A database of active incidents, work zones, and toll facilities and rates. 3. Speed data from INRIX (separate from the Trips data set). Tools and Analysis Description The MDOT SHA will leverage a “Route Analytics” tool embedded within the RITIS Platform. This tool will tell the MDOT SHA which routes travelers are using to detour around an incident, including the number of trips that took the detour. It will also provide the travel times (average, minimum, or maximum) and reliability of each route. The MDOT SHA will explore locations that have higher crash rates to attempt to understand travel behavior, especially for daily com- muters. The goal is to understand if there are vulnerable neighborhoods or arterials that are negatively affected by excessive detours, and if so, to come up with a plan of action to minimize the impacts. This could be in the form of detour restrictions, better information campaigns, partnerships with third-party navigation apps, or enforcement. Once enough data have been collected and analyzed, the intent is to implement real-time modeling and decision support based on historical data and to use this information in combi- nation with real-time monitoring to predict behavior and proactively manage work-zone and incident detours to minimize impacts. Results and Outputs The anticipated results of this approach include predicted conditions in and around a work zone that will be activated in the future, and the ability to export that information to third-party applications such as Waze, Google Maps, RITIS, and others. This information can then be used by third parties to better inform travelers and encourage change in routing. As suggested in Figure 45, the MDOT SHA also anticipates being able to evaluate permit requests and activations and use modeling to predict impacts and explore alternative options that may result in a lower impact to travelers by shifting work-zone times or closures. For example, vehicle probe data can be used to calculate the expected LOS at different times and to provide recommendations for permit parameters. Similarly, predictions can be used to adjust operational response plans and pre-stage ITS devices based on expected behavior. For example, work-zone relevant messaging can be posted to DMS upstream from the work zone depending on anticipated queues and potential detour routes.

62 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Intersection Performance Analysis: Percent Arrivals on Green and Turning Movements (Austin, TX) Overview Automated Traffic Signal Performance Measures (ATSPM) is a suite of performance measures, data collection, and other related analysis tools that support the management of traffic signals. ATSPM is widely deployed in both Utah (https://ops.fhwa.dot.gov/publications/fhwahop18048/ index.htm) and Georgia (https://ops.fhwa.dot.gov/publications/fhwahop18050/index.htm), and is gaining popularity around the country because of the insights that it provides agencies with respect to measuring percent arrival on green and other signal metrics that provide actionable insights into the performance of a signalized intersection. However, the costs to deploy ATSPM at intersections can be high due to the types of sensors and communication infrastructure that must be installed at each intersection, and the size of the data coming out of each intersection can be expensive to manage. The city of Austin is implementing a cheaper alternative that leverages speed and location data sourced from connected vehicles in a way that mimics some of the ATSPM metrics, but Figure 45. Ranked list of detour routes used to go between I-495 and a northern section of I-95 (Source: MDOT SHA Trips Analytics Suite).

Case Examples 63 without the need to deploy any roadside infrastructure or communications equipment. The goal of their deployment is to achieve ATSPM-like capabilities, but at a substantially reduced cost that is also significantly more scalable. Data Source Description The following data sets were used in this development effort: 1. Shapefiles of the location of individual signalized intersections. 2. Individual waypoints (or bread crumb trails) of vehicles moving throughout a network. Data are sourced from connected vehicles that are typically comprised of a sample of the traffic stream. Because actual vehicle positions are known every few seconds, the precise speeds and pathways through signalized intersections can be isolated and analyzed. Tools and Analysis Description The system Austin has deployed is from INRIX and is based off of research from Wayne State University and the University of Maryland CATT Lab. Vehicle movements (location, speed, and heading) can be plotted through individual intersections. Figure 46 is a time/distance chart that shows typical approach speed, travel time through an intersection, and two other measurements needed to compute control delay, the reference travel time and the delay. The slope of the dots in the graphic represents the speed. A horizontal slope would indicate a full stop, and an almost horizontal slope indicates a rolling stop while approaching the intersection. The researchers developed a system that allows agencies to select a set of intersections to analyze. As shown in Figure 47, date ranges and time periods are entered, and the results come back in table and graphic/chart formats. Figure 46. The y-axis represents distance, and the x-axis represents time. The dots are the location of a single vehicle as it traverses the corridor. Changes in slope represent slow-downs or delays at an intersection (Source: INRIX).

64 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Figure 47. The user selects a series of signals (pink dots) to analyze for a specific date range, time of day, and day of week (Source: UMD CATT Lab). Results are returned in three interactive panels on one screen, as shown in Figure 48. The “Ranking and Summary Table” is found across the top, the “Intersection Selection Map” is at the lower left, and the “Intersection Data Display” is at the lower right. Making selections in any of these panels will automatically be reflected, as appropriate, in the other two. Every movement of every intersection in the user’s query, commonly 12 movements per inter- section, has a row in this table. The ranking of the movements comes first. The user decides which metric is the basis for ranking by clicking the desired header to the right. The next three columns define the intersection, approach and movement. The metrics further right are displayed based on user preference, indicated by checking boxes in the “Display Options” menu at the top right. The outputs of the system include the following: • Travel time (average and maximum): these are based on the time it takes for each sampled vehicle to traverse the 230-meter analysis zone representing the intersection. • Approach speed (average and maximum): these are based on the highest calculated speed of each sampled vehicle (based on two consecutive “pings”) while traversing the 150-meter approach zone. • Control delay (average and maximum): for each sampled vehicle, these are based on the differ- ence between the travel time and the reference travel time. The latter is the 5th-percentile fastest travel time recorded for each movement of the intersection, and is intended to represent what the travel time would be in the absence of any intersection delay. • Percent (arrival) on green: this is the percent of the sampled vehicles that proceeded through the intersection without stopping (as defined above). Mathematically, it is (a) the difference between the vehicle count and the stopped vehicle count; divided by (b) the vehicle count; and then (c) reported as a percentage.

Case Examples 65 Figure 48. The results are displayed as a table of metrics (top), a map showing the highlighted intersection (bottom left), and an intersection diagram (bottom right) that depicts turning movement delay, percent arrival on green, vehicle counts, and intersection orientation. (Source: UMD CATT Lab). Results and Outputs It is too early in the deployment of this technology to have documented results. Thus far, deployment costs are lower than traditional ATSPM, and $50K per year will typically cover ∼100 intersections. The capabilities of Austin’s implementation are not an exact match to that of ATSPM. There are certain things that ATSPM can do that the connected vehicles data cannot yet do because of sample sizes, and the non-real-time nature of the data analysis. The city will be evaluating the results of the deployment over the coming year. Analyzing People Movement Before and During the COVID-19 Pandemic (U.S. DOT, MDOT, and MTI) Overview The U.S. DOT is seeking to better understand the impacts of COVID-19 travel restrictions on mobility and safety in the United States. They are working directly with the Maryland Trans- portation Institute (MTI) to measure rural and urban speed changes on the entire NHS weekly from January 2020, through the end of the calendar year. They are also quantifying the number of individuals who are traveling in the United States on a daily basis. Statistics that are sought by the U.S. DOT include the number of trips per county broken down into the following distance bins: • < 1 mile is around home or neighborhood. • 1–3 miles is surrounding neighborhoods that can be reached by walking. • 3–5 miles reflects a small city or suburban complex. • 6–10 miles reflects citywide travel. • 11–25 miles reflects small metropolitan complexes. • 26–50 miles reflects big metropolitan complexes. • 51–100 miles is the transition between local and long distance.

66 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation • 101–250 miles is a moderate 1-day drive or a good range for rail; some flying trips. • 251–500 miles is a long 1-day drive or rail trip; flying more likely. • > 500 miles is beyond a 1-day drive; flying much more likely. Data Source Description To analyze changes in average speeds on the nation’s roads, the U.S. DOT is leveraging the NPMRDS probe-based speed data. These data were leveraged because it is readily available and already paid for by the U.S. DOT. To analyze the number of trips taken per day broken down by distance bins, the U.S. DOT is leveraging LBS data from several commercial providers. LBS data are one of the only data sources that can be collected nationally covering a large percentage of individuals. Existing algo- rithms had already been developed that could also compute metrics from these data sources. Tools and Analysis Description The U.S. DOT contracted with the University of Maryland MTI for data processing and analysis. The MTI team ingested LBS data from multiple providers, cleaned the data, fused the LBS data with other geospatial data (like road networks), and used multiple artificial intelligence techniques, cluster analysis, and neural networks to impute various metrics, as shown in Figure 49. Figure 49. The method used for computing various statistics for the COVID-19 mobility impacts platform. More information is available at https://data.covid. umd.edu/.

Case Examples 67 These data are provided to the U.S. DOT via CSV and Excel files. However, MTI has also built an online impact analysis platform to make it easier for users to explore various mobility, economic, health, and safety data, as shown in Figure 50. Results and Outputs In addition to the requested travel distance bins, the MTI platform also visualizes daily stats at the county, state, and national level regarding the number of trips taken, trip purpose, trip mode, economic impact data, health data, and other demographic data and indices. The data were used to analyze the immediate shift in travel after the state of Georgia decided to partially reopen on April 24, 2020, ahead of most of the country and surrounding states. The day of the opening, the analysis of the GPS data showed that the percent of individuals staying home fell by 32%. The distance traveled per person went up by 19%. The number of out-of-state trips to Georgia increased by 13%, and many of those trips came from neighboring states that had yet to reopen, as shown in Table 9. The MTI team has also attempted to summarize certain statistics to draw conclusions that help the U.S. DOT and other decision makers understand the impacts of the travel restrictions, executive orders, and public adherence. Figure 51 and Figure 52 show the results of some of these analyses. The MTI team will continue to deliver data to the U.S. DOT on daily trip statistics at least through 2021. The LBS data were also useful in analyzing movement and the correlation of new cases with notable mobility behavior trends. For example, the Pleasant View Nursing Home in Maryland saw the greatest number of concentrated deaths in the state; 81 residents and 36 staff members tested positive for the virus and 24 residents died from the disease. The LBS data were used to develop a community-level contact tracing application that showed travel between the facility and other hot spots in the state. Figure 50. Screenshot of the MTI COVID-19 Impact Analysis Platform available at data.covid.umd.edu.

Table 9. Travel increases measured the day after Georgia’s partial reopening on April 24, 2020 (Source: data.covid.umd.edu). Figure 51. The measured impact of stay-at-home orders on travel behavior (Source: https://data.covid.umd. edu/findings/index.html).

Case Examples 69 Figure 52. Derived social distancing index by state and time (Source: https://data.covid.umd.edu/ findings/index.html).

Next: Chapter 5 - Summary of Findings »
Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation Get This Book
×
 Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation
Buy Paperback | $79.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Over the last decade, state departments of transportation (DOTs) have begun to use vehicle probe and cellular GPS data for a variety of purposes, including real-time traffic and incident monitoring, highway condition, and travel demand management. DOTs are also using vehicle probe and cellular GPS data to inform system planning and investment decisions.

The TRB National Cooperative Highway Research Program's NCHRP Synthesis 561: Use of Vehicle Probe and Cellular GPS Data by State Departments of Transportation documents how DOTs are applying vehicle probe and cellular GPS data for planning and real-time traffic and incident monitoring and communication.

In December 2021, an erratum was issued.

READ FREE ONLINE

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!