The Segment Analysis: A Better Way to Measure Corridor Travel Behavior with StreetLight InSight®

We’ve added a brand-new type of analysis to StreetLight InSight®: the Segment Analysis. This feature helps you measure corridor travel behavior faster and more comprehensively. In this blog post, we’ll show you how the feature works and three great ways to use it:

• Diagnosing the Cause of Congestion
• Multimodal Planning
• Before-and-After Studies

What Is the Segment Analysis?

The Segment Analysis Project Type provides information about trips through corridors. The analytics included in this Project Type are derived from Navigation-GPS data, which are location records created by connected cars and turn-by-turn navigation apps for mobile devices.

The following standard analytics are provided for every Day Part in your study, and you can customize these Day Parts down to one-hour intervals:

• Trip speed
• Trip duration
• Relative volume of trips

When we calculate average speed along the segment, we look at the travel time between gates – that is the information our clients are usually looking for. But we’re going beyond these standard Metrics with this new Project Type.

The Segment Analysis comes with an important new Metric: Free Flow Factor, which you can also think of as “Speed-As-Share-of-Free-Flow.” It is a ratio of the average trip speed for a Day Part to the maximum average trip speed in any hour during the entire Data Period.

This means that when the Free Flow Factor is close to 1, there is little congestion on the corridor. The lower the Free Flow Factor, the greater the congestion. However, the “maximum average speed” per hour has to have at least 5 samples to prevent one or two outlier “speed demons:” from skewing the results.

How Does the Segment Analysis Help With Corridor Studies?

Some of you may be saying, “Hey – I’ve actually used StreetLight InSight” to get this information in the past. This isn’t new.” But the key point is that our Segment Analysis makes it so much easier to get this information. It’s faster and quicker to set up your study, and you don’t have to manually process tons of Metrics in Excel to get the insights you need.

The first step to using our Segment Analysis is to set up your Road Segment Zones, and we’ve made it really easy to use your own line geometries. All you have to do to set up your Zones is to upload a line segment shapefile with all of the corridors that you want to analyze. Road Segment Zones uploaded via line segment shapefile are the only type of Zone that can be used for this Project Type.

Once you upload your Line Segment Zones, three gates will automatically be created: one at the beginning of each segment, one at the middle of each segment, and one at the end of each segment. (See Figure 1 below).

Figure 1: This shows a Zone created by uploading a line segment shapefile to StreetLight InSight. Three gates, shown in light blue, were automatically created: one at the beginning, one at the middle, and one at the end of the segment.

Once your Zone Sets are created, simply navigate over to the Create Projects tab, then create your Segment Analysis Project just like you would any other Project. You customize your Day Parts using our standard Project Options menu, and you can set up your bins for Trip Speed and Trip Duration in separate Project Options menu. It’s that easy.

Example #1: Diagnosing the Cause of Congestion

Many cities already know where their congestion occurs thanks to a variety of different tools — including personal misery on the way home from work! The new Segment Analysis Project Type offers a way to explore when and why congestion happens, especially when used in combination with Premium Trip Attributes, Zone Activity Analysis, and/or StreetLight Volume: 2016 AADT.

Of course, the Segment Analysis can also help you identify where congestion occurs, and that matters. In our experience, our instincts about where and when the worst traffic happens are not always correct. Being stuck on your way to an important meeting may feel a lot more stressful than a slowdown during lunch hour.

As a quick example, let’s analyze a few roads around Los Angeles, which is famous for its mind-numbing congestion. Imagine that LA residents are concerned about congestion on Del Mar, so planners in LA want to identify the time and place of top congestion, and identify why it is occurring. They’ll use this information to implement changes, and, later, to see if congestion improved.

Figure 2: These are our roads of interest around Vista Del Mar in Los Angeles.

First, we can use the Segment Analysis to compare average speed to volume of traffic. As Figure 3 (below) shows, during the afternoon rush, Vista Del Mar’s speed is actually quite high compared to other roads in the area. It’s comparable to Imperial Highway and a little faster than parallel Pershing and Sepulveda. This is true despite the fact that Vista Del Mar moves more people than those other two roads, thus its throughput is in fact quite good – at least it is on an LA-adjusted basis.

Figure 3 – This shows average trip speed compared to total traffic for various roads around Vista Del Mar in LA. We created this chart in MS Excel using StreetLight InSight .CSV files.

Next, we can use another type of Project – Origin-Destination to Pre-set Geography Analysis – to dig into the origin and destination of traffic during rush hour on Vista Del Mar. This will help us to better understand the causes of existing traffic, and then target solutions to the locations indicated.

Figure 4 – This shows the destinations of trips that passed through the congested segment of Vista Del Mar. “Hotter” zones can be good targets for transit, employee incentives, rerouting, and other programs to reduce congestion.

Example 2: Multimodal Planning

One of the biggest benefits of Big Data is the ability to scan really large geographies for specific travel conditions. This is particularly helpful for multimodal planning, and the Segment Analysis Project Type gives our clients a new way to do this.

Take, for example, bike lanes. Let’s imagine we want to do in-depth studies to invest in bike lanes in Richmond, Virginia. Before we spend a lot of time and money on those studies, we want a “short list” of the roads that are the best candidates for bike lanes. Let’s hypothesize that the best place to add bike lanes are road segments where:

• There are lots of short vehicle trips happening in cars.
• Those vehicle trips would be faster if they were bike trips instead.
• The average speed of cars and trucks is less than 40mph – this makes cycling safer for riders.

In the pre-Big Data world, planners could look only at speed limits, some level-of-service data for vehicular travel, and perhaps some surveys. That data was often inaccurate, took a while to collect, and was hard to compile. Now with Big Data and our new Segment Analysis, we can look at all three of these characteristics simultaneously. In fact, by combining the Segment Analysis Project Type with StreetLight’s existing Premium Trip attributes, we can get everything we need to complete this study in under 5 minutes.

First, we create our Zone Set by uploading Line Segment shapefiles to StreetLight InSight. Next, we run two studies: Segment Analysis and Zone Analysis with Premium Trip attributes. (I won’t go into the details here, but please check out our Support Center for more information).

You can see the results of the Segment Analysis as displayed in StreetLight InSight in Figure 5 below.

Figure 5: This map shows the congestion factor for candidate bike lane roads. The roads that are light green and yellow are more congested for personal vehicle traffic, making them potentially better candidates for biking, as drivers might see more time saving benefits by switching modes.

My next step is to assign five “scores” for each of my road segments – congestion during rush hours, max daytime hourly speed for cars, max daytime hourly speed for trucks, share of personal car trips under 5 miles. I eliminate road segments that had one or more scores “out of range” (speeds over 40mph, fewer than 5% of trips under 5 miles, etc.).

This leaves me with a shortlist of 10 or so roads to pursue in greater detail (shown in Figure 6 below). I can now use local knowledge, stakeholder outreach, and all the other planning tools in my toolkit to finish my multimodal planning process, but I have a huge head start thanks to Big Data.

Figure 6: Top 10 road segments for bike-ability, based on criteria from StreetLight Segment Analysis and Trip Attributes analysis.

For an example of how these “rankings” work in chart format as opposed to map format, take a look at our blog post on electric vehicle charging stations.

Before and After Studies to Measure Changes in Speed and Delay

Our new Segment Analysis also makes measuring the impact of infrastructure on speed and delay much faster and easier. It’s a quick-and-easy check on how speed and congestion change over time on a particular road segment.

To demonstrate this, let’s take a road in Maine where signal timing improvements were implemented in February 2017. We measure the speed on the segment and subsegments of the road in question for six months before and six months after the implementation of the timing improvements. Our question is: Did the timing work? It did indeed! In particular, travel speed improved in the central two segments.

Figure 7: Four segments near improved signals in Brunswick Maine.

As shown in the table below, gains in the central segments were significant: 5 to 6mph or about 20%.

Analyzing Roadways with the Segment Analysis: Putting it All Together

As you can see, our new Segment Analysis is a powerful way to access actionable insights about trips through corridors. It makes collecting all the pieces of information you need to understand – and solve – congestion challenges much simpler and quicker.