Navjoy Helps CDOT Track Travel Time Reliability

By Justin Healey, Regional Business Development Manager

When it comes to travel time, people care more about reliability than how long it takes to get from A to B. For example, most people would prefer a consistent 45-minute commute to work each day over a commute that varies anywhere from 20 minutes to an hour. That's why it's important to develop a system to analyze and monitor travel time reliability.

Traffic near Larimer Square in Denver, Colorado.The Colorado Department of Transportation (CDOT) hired Navjoy to develop a Travel Time Reliability Monitoring System (TTRMS) using funding from the Strategic Highway Research Program 2 (SHRP2). Development of the TTRMS implements two products from the SHRP2 Reliability Data and Analysis Tools bundle. This post provides a summary overview of Navjoy’s ongoing work to help CDOT understand the relationship of various sources of congestion on travel time reliability.

A TTRMS draws correlations between travel time data and one or more sources of non-recurring congestion, such as inclement weather, a crash or stalled vehicle, a special event, or a work zone. The tool then allows the user to visualize which incidents affect the reliability most.

CDOT's need

CDOT was well-positioned to implement a TTRMS because it gathers robust traffic volume, speed, road closure, incident, and travel time data from a variety of systems monitoring highways state-wide. Real-time data is processed and fed into databases from which historical information can easily be retrieved and assessed.

Furthermore, there were two driving forces that increased CDOT's need for a TTRMS:

Car locations being tracked at an intersection1. Topography and Freight Movement – Colorado's mountainous terrain limits thenumber of alternate routes available on many main highways. This was illustrated recently with the closure of I-70 through Glenwood Canyon. Additionally, some interstates and state highways with limited alternate routes are high-priority freight corridors, so delays are more costly. Incidents, severe weather, special events and road work can all periodically cause significant delays on these roadways.

2. Data Silos – CDOT has an extensive network of sensors to collect data, but it's often segregated in different systems. Essentially, a TTRMS enhances CDOT’s understanding of congestion impacts and allows them to strategically work to improve the collective reliability of Colorado’s transportation network.

TTRMS Development

Initial development of the TTRMS began by exploring work zone data from one interstate corridor in the Denver Metro area. The pilot corridor allowed our project team to understand data trends within the sources and review preliminary results to identify desirable features for a tool with more processing power. This produced a TTRMS with three key components:

1. Data cleaner – When too much raw travel time data is available, it can be daunting. The team needed an easy way to process and clean large amounts of raw travel time data before merging it with impact data. This was necessary to ensure a consistent data format between the non-impacted and impacted travel time data. Within the data cleaner, impacts are classified by the major type (incident, work zone, etc.) and the sub-type(s) (e.g., work zone – fiber installation). Cleaned data are inspected, then uploaded to the database.

Data and graphs overlaid on a city skyline2. Database queries – The database stores cleaned data and allows users to easily assess and compare travel time impacts across different incident types (weather, work zones, etc.) and corridor types (urban interstate, rural highway, etc.). A TTRMS tool extracts numbers from the database and places them into a spreadsheet for analysis. The returned data can also be limited by selected criteria.

3. Graph generator – Lastly, the database-generated spreadsheet goes to the graph generator, which creates interactive charts to summarize impacts. Violin plots and Cumulative Density Functions (CDFs) are generated to show the distribution of travel time impacts according to any criteria selected.

The tool allows users to interact with their data to conduct high-level, detailed comparisons from the TTRMS.

What's next?

An illustration of dataNavjoy will continue to work with CDOT to enhance the tool throughout the remainder of the project. We intend to add functionality to evaluate impacts from simultaneous events (e.g., work zone and incident), as well as a database management interface tool. Additionally, the database will continue to be built up by cleaning and uploading data from more incidents. Finally, we will keep refining the user interface to make the tool easier for users to glean insights that can lead to actionable implementation use cases (e.g., top locations to deploy additional courtesy patrol vehicles).

The completion of the TTRMS coincides with another significant project being undertaken by CDOT. To consolidate all of their data sources, CDOT is developing a real-time data hub that will process and store data from all sensors. As a research project, the TTRMS established a methodology that can be easily replicated and implemented on a much larger scale for analyzing the travel time reliability impacts of non-recurring congestion. Navjoy is currently working with CDOT to integrate the TTRMS data model into the real-time data hub.