Unleashing the Power of Connected Vehicle Data
PACCT is a fully scalable web-based analytics platform for connected car trajectories. The platform can ingest, warehouse, analyze, and visualize billions of connected vehicle datapoints per month. The platform utilizes the latest cloud and distributed computing technologies to dynamically scale the infrastructure based on the cyclical load of datapoints that are received from Wejo.
Transportation Operations Staff can use PACCT to monitor roadway conditions in near real-time and Planners can use it to support data-driven decision making by analyzing recurring congestion or quantifying the effects of roadway improvements.
Reducing Traffic Congestion with Data-Driven Solutions
In today’s world, traffic congestion has a direct impact on the economy. Traffic delays on already overcrowded roadways can occur in a matter of seconds. Traffic Management Centers need more than Dynamic Message Boards and traveler information websites to address recurring and non-recurring congestion. Traffic Planners need constantly updating data as travel patterns change monthly or even daily as motorists adapt to ever changing roadway conditions.
PACCT offers users the ability to make data-driven decisions by leveraging connected vehicle data that has unparalleled accuracy and granularity with low latency. Each of the tools in PACCT was developed with engineers, planners, and operations staff use cases’ in mind to distill the information in a user-friendly and intuitive manner
We developed an analytics infrastructure to ingest, process, and render CV data in near real-time. This means that on a daily basis in the Tri-State Region (NY, NJ, CT), we received and processed approximately:
Datapoints
1.2B
Per Day
Journeys
3.5M
Per Day
Database Size
195
GB Per Day
Datapoints
334M
Per Peak Period
Journeys
1.1M
Per Peak Period
Database Size
65
Per Peak Period
Vehicle Movement Dataset
The Wejo Vehicle Movement Dataset that is used in PACCT is sent directly from the vehicles, has an update interval of 3 seconds, and latency under a minute. Some characteristics of this data include:
- CV Data is precise to +/- 10 feet (under 3 meters) 95% of the time. This can help to identify individual lanes of a highway, back of queue, or even parking spots.
- The Mobility Dataset differs from probe data and other data sources because it is transmitted directly from the vehicle itself. It is more precise, updated more frequently, and has significantly less latency than traditional datasets.
- Data from Connected Vehicles covers better than 95% of roadways in the United States and transmits 24 hours a day, 365 days a year.
Realtime Conflation and Live Speed Layer
Infosenseglobal & KISNN has developed a proprietary conflation architecture to connect the vehicle datapoints to a linkset in real-time. This unique process allows data to be usable for analysis in ~15 seconds. Origin/Destination insights are available in ~5 minutes after the trip ends.
- With this architecture, users can view trip & O/D analytics at the link level, utilize particular links in a pass-through query, and calculate link-based speeds for the live speed layer.
- Operators are able to monitor the roadway network in near-real time and make effective decisions based on the latest available data.
- The Live Speed Layer can also be integrated into open source maps, like OpenStreetMap, or paid mapping services like Google Maps.
Origin and Destination (O/D)
Insights into the origins and destinations of trips are another benefit of the Wejo CV data. Because of the high degree of location accuracy +/-10 feet (under 3 meters) we are able to accurately determine the origin and destination of a vehicle, as well as the precise route that vehicle took.
- Each trip is assigned an anonymized unique identifier that allows us to determine the origin and destination.
- Each GPS point contains geolocation information, direction of travel, and speed. This information is parsed and connected to previous GPS points per trip to create the vehicle’s route.
- Multiple spatial aggregations are available for analysis in PACCT. These include county, town, or point of interest levels as well as an option to upload custom zone shapefiles.
Driver Event Dataset
The Wejo Driver Event Dataset includes information on when/where the vehicle was started/stopped (ignition on/off), exterior temperature, seatbelt status change, windshield wiper status, harsh braking events, and rapid acceleration events.
- The seat belt status, harsh braking, and rapid acceleration events data can be used to determine hazardous roadway conditions or identify potential safety issues.
- Exterior temperature or windshield wiper status can provide explanatory power for harsh braking events during a weather event. The data elements can be used together in this manner to create greater insights.
Data Privacy
Data privacy is a primary consideration when working with CV data. The raw data is anonymized and does not include any Personally Identifiable Information (PII). Beyond this, the architecture was developed with business rules to enhance privacy.