Three years ago, INRIX introduced INRIX Trips, also known as INRIX Trip Reports. INRIX Trip Reports provides accurate insight into the trips people take, including where they begin and end their journeys and all the waypoints in between. There are three types of points in the INRIX Trip Reports data; origin points, waypoints and destination points. Trips start at an origin point, generate data at a number of waypoints and end at a destination point. The INRIX Trip Reports data can be used to analyze movement patterns, generate origin/destination matrices and perform select link analysis in tools such as DB4IoT with INRIX Trips.
INRIX is now introducing INRIX Trip Paths. In March INRIX provided us with an INRIX Trip Reports dataset for a project with DKS Associates and David Evans Associates for the Regional Transportation Council (RTC) and the Washington State Department of Transportation (WSDOT) in Clark County (WA). INRIX also provided us with an early version of INRIX Trip Paths for the same area and the same time period. We merged the two databases in DB4IoT. This provided a rich dataset for RTC and WSDOT. As a side benefit it also provided us the opportunity to compare the two products in detail before INRIX Trip Paths was even officially announced by INRIX.
Read the INRIX Trip Paths launch press release.
Learn more about DB4IoT, INRIX Trip Paths and how DKS used both in their recent project for RTC at the following events:
Exhibits:
June 5-7, ITS America, Washington DC, Moonshadow Mobile in Booth #754
June 3-6, ITS Europe, Eindhoven, The Netherlands, INRIX in Stand 0.23
Conference Presentations:
June 4, TRB AppCon, Portland, OR (Moonshadow/DKS)
June 22, OMUG, Salem, OR (RTC/Moonshadow)
June 23, ITE Western District, Monterrey, CA (DKS)
Selecting Trips in INRIX Trip Reports Using Waypoints
Waypoints are generated when a vehicle sends a signal to the car manufacturer with its position, speed and a number of other variables. The frequency at which vehicles send their location varies. Some vehicles update their location every ten or twenty seconds while others may report only once every few minutes. If a vehicle reports its location once per minute and a road segment such as an off-ramp takes only 15 seconds to drive, then only one in four vehicles that take the off-ramp will generate a waypoint close to the off-ramp. This means that if you select all vehicles on an off-ramp by placing a polygon around the off-ramp you will only find one in four relevant trips. As we all know, GPS has limited precision so a portion of the relevant waypoints will lie further away from the off-ramp and it can be difficult to select all waypoints on the off-ramp and be sure you are not including waypoints that are actually on another road segment.
These problems are alleviated by looking at larger sample sizes. Instead of looking at peak traffic on weekdays for a month we can look at a three or six-month time frame to make the data more representative. We usually bring up datasets with INRIX Trip Reports that contain hundreds of millions of waypoints representing millions of trips. In the tests we conducted we imported all trips that had at least one waypoint in Clark County (WA) in 2018. This data set contains 380 million waypoints representing five million trips.
Figure 1. Waypoints in DB4IoT with INRIX Trip Reports
Figure 2. Select trips going through an off ramp with a polygon filter
Figure 1 shows all waypoints on the cloverleaf intersection of I-205 with SR-500 in Clark County (WA). In Figure 2 we selected all waypoints on the I-205 northbound exit onto SR-500 westbound by placing a polygon around the exit. By selecting the waypoints we have identified all the trips that go through this off-ramp. Since we know the trips we can look at their origins and destinations. In Figure 3 we show an Origin-Destination Matrix for all traffic going through the off-ramp.
Figure 3. O/D matrix by city using waypoint polygon selection in INRIX Trip Reports
Data shown is for demonstration purposes only
Selecting Trips in INRIX Trip Paths Using Road Segments
The INRIX Trip Reports and INRIX Trip Paths data is derived from the same source information but the data looks very different. Whereas INRIX Trip Reports contains waypoint location points for trips, INRIX Trip Paths contains a list of road segment IDs for these same trips.
The INRIX Trip Paths data is created in the following way. The waypoints for each trip are first mapped onto road segments in OpenStreetMaps (OSM). INRIX then determines which road segments are skipped. If a vehicle has generated a waypoint on SegmentID A and the next waypoint is on SegmentID E then INRIX determines the most likely path and also assigns SegmentIDs B, C, and D to this trip and makes an estimate of the time at which each road segment was entered and exited. INRIX Trip Paths simply includes a list of OSM SegmentIDs for every trip.
Figure 4. INRIX Trip Paths stores road segments for trips
Because we now have lines that represent the road segments, selecting all trips that go through this off-ramp becomes much easier. We can simply click on the off-ramp and add it to our filters. The number of trips that are selected also increases dramatically, by an average 2.2 times according to our tests. The shorter the road segment the larger the increase. This means that if you generate an Origin/Destination matrix for all trips going through a short road segment you will get a higher number of results in DB4IoT with INRIX Trip Paths versus DB4IoT with INRIX Trip Reports. In Figure 5 we have selected the same off-ramp as in the example where we used INRIX Trip Reports but in this case we selected the off-ramp by filtering by its segment ID.
Figure 5. Select trips going through an off ramp by clicking on the off-ramp
When you zoom out in DB4IoT with INRIX Trip Paths you will see how the traffic from the off-ramp is dispersed to different roads that it feeds into.
Figure 6. Select link analysis using INRIX Trip Paths
Figure 7. O/D matrix by city using road segment selection in INRIX Trip Paths
Data shown is for demonstration purposes only
When we generate an O/D matrix for eight cities for the traffic going through this off ramp we have a total of 18,327 results in 2018 in INRIX Trip Paths . When selecting trips with a polygon filter on the waypoints in INRIX Trip Reports we found 8,449 trips in the data. On average INRIX Trip Paths yields 2.2 times more results than INRIX Trip Reports. As always, the numbers are not statistically significant for low-count O/D combinations using this off-ramp.
Figure 8. Increase of number of trips found in INRIX Trip Paths versus INRIX Trip Reports
Data shown is for demonstration purposes only
RTC & WSDOT Use DB4IoT with INRIX Trip Reports and INRIX Trip Paths
On May 14, 2019 RTC organized a day-long workshop to brainstorm about corridor improvements in Clark County (WA). DKS Associates and David Evans Associates presented their findings to transportation engineers from RTC, WSDOT, ODOT, Clark County and the City of Vancouver.
At this event DKS Associates used DB4IoT with the merged Trip Reports and Trip Paths data to answer questions that came up during the brainstorming sessions. RTC wanted to know how much traffic is leaving the freeway north of Vancouver and driving through downtown to avoid congestion on I5 Southbound during the morning peak. The INRIX Trip Reports data showed the actual locations of vehicles in downtown Vancouver that took an I5 off-ramp north of Vancouver only to enter the freeway again south of Vancouver.
Figure 9. DB4IoT with INRIX Trip Reports and Trip Paths is used in an interactive workshop
“Here at the SW Washington Regional Transportation Council (RTC), we are responsible for dealing with regional transportation planning for the Vancouver, WA metropolitan area. We’ve been working with DKS Associates for a number of years, and are very excited about the valuable and innovative tool they have built in partnership with Moonshadow Mobile using INRIX Connected Vehicle data. The dynamic Origin and Destination analysis tools allow us to easily ask questions around regional trip distribution, complex trip corridor patterns, and diversion on local streets. The dashboard approach to data analytics allows us to ask questions on the fly, in a workshop setting, and to answer new questions as they come up.”
On I205 NB there are two on-ramps that can be used for traffic coming from NE Fourth Plain Blvd. One of the on-ramps is congested whereas the other is underutilized. WSDOT wanted to know how much traffic from NE Fourth Plain was taking the congested on-ramp. Using data from INRIX Trip Paths DKS was able to show that 28% of the traffic on the congested on-ramp would be better off taking the underutilized on-ramp.
Figure 10. Analyzing an on-ramp in DB4IoTwith INRIX Trip Paths
Anonymization
Traffic engineers and planners look at large-scale movement patterns of vehicles. They do not need or want personally-identifiable information about drivers to do their work. Neither do OEMs want to distribute any identifiable information about their drivers.
The data in INRIX Trip Reports is anonymized. When we receive the data all identifying information has been removed. We do not get the vehicle make or model. We also do not get the exact location of the origin and destination so we can’t trace trips back to addresses. What we do get, however, is an exact timestamp and a GPS location for each waypoint on the trip. While it is impossible to track the trip to a specific person or vehicle you do have information on the exact time a vehicle was at a certain location.
INRIX Trip Paths goes one step further in anonymizing the data. All we receive in INRIX Trip Paths is a list of road segments with their entry and exit times. These times are interpolated from the waypoint data so we no longer have an exact time and location for the vehicles. For generating Origin/destination matrices or performing select link analysis this makes no difference. The large-scale conclusions you can derive from the data remain the same.
Combining INRIX Trip Reports and Trajectories
There is one drawback when using INRIX Trip Paths instead of INRIX Trip Reports; you do not have access to the original source data anymore. INRIX Trip Reports data sets usually include hundreds of millions of waypoints collected from moving vehicles. Data at this scale always contains some errors or data that you would want to exclude. If you have the original source data then you can make better judgements about the bias in the data and records you may want to exclude for a certain type of analyses.
With INRIX Trip Paths you no longer have access to the original waypoint data and therefore rely on INRIX to have filtered out the incorrect data. We imported both the INRIX Trip Reports and INRIX Trip Paths data into DB4IoT and this provides the best of both worlds. Whereas INRIX Trip Reports works with points (waypoints), INRIX Trip Paths works with lines (road segments) so we had to make these two very different datasets compatible in order to merge them. To do this we generated a point for every trip on every segment. As a result you can do queries in DB4IoT that seamlessly combine results from both INRIX Trip Reports and INRIX Trip Paths data. Of course this dataset is much larger than either the INRIX Trip Reports or INRIX Trip Paths data alone. One year of Clark County data increased from 380 million records to well over one billion records.
Conclusions
INRIX Trip Paths is a major improvement by INRIX for drawing conclusions from the trips data. It provides a two-times improvement on the INRIX Trip Reports data for generating O/D matrices and performing select link analysis. The trips data is further anonymized in INRIX Trip Paths and it is easier to make selections. The only drawback is that you do not have direct access to the original waypoints but that drawback can be overcome by combining the two data sets for use cases that need this.
