TrafficCast Q&A with GPSLodge
I sat down for a Question and Answer session with the CEO of TrafficCast, Connie Li. TrafficCast has developed an algorithm to aggregate not only traffic data but data impacting traffic, like weather and events. We all know that heavy traffic and accidents slow your drive, but so does rain and snow.
As we move forward, and more people have and use smart navigation devices, the ability to tie up all kinds of data into usable information streams will play a big part in our lives. TrafficCast is getting a jump on reliable traffic information gathering and on gathering all that other information too.
Q&A After the jump...
GPSLodge - Can you tell me about TrafficCast, how did it get started?
Connie Li - Bin Ran and I founded TrafficCast in 1997 and in 1999 delivered the first online traffic data. As a Professor and the former Director of the Traffic Lab at the University of Wisconsin at Madison, Bin has a special interest in this field, along with myself. Before TrafficCast, I founded a traffic and transportation consulting business, TranSmart Technologies, and in that work I began to see a pattern in client requests and consulting work that indicated a need for an increasingly disciplined approach to traffic data analysis. Bin & I worked to apply his research in this area, and we decided to start a new chapter in our careers to launch a traffic information company grounded in advanced traffic flow theory and traffic science.
GPSLodge - TrafficCast covers hundreds of thousands of miles with traffic reporting; what technologies do you use to accomplish that?
Connie Li - Fundamentally, we believe that accurate traffic information and travel time forecasts require analysis of the range of factors that impact traffic. TrafficCast collects real-time information from more than 350 data sources, which we believe is more data points than anyone in the market, to produce accurate real-time traffic information. These data points include weather, construction, incidents and sensors. We built a sophisticated software architecture to assure access to these data sources, some of which are based on our various partnerships and alliances with Westwood One, MyWeather, Weather Central, Speed Info and ESRI.
GPSLodge - People are pretty familiar with traffic sensors, but how do you incorporate Cellular Phone data to help understand traffic conditions?
Connie Li - The anonymous mobile phone position and signaling data in a GSM Network are collected, analyzed and converted into travel time and speed information for major highways and surface streets. Currently in Shanghai, the system collects cellular data from 14 million China Mobile subscribers and covers all roads within the Inner Ring Area, including 483 miles of major roads comprising 1,700 roadway segments.
We've seen that such data can improve model accuracy. Beyond cell probes, we're also working with a range of floating vehicle data (FVD) collected using GPS devices, Wi-Fi Hot Spot technology and cellular probe technology. For speed and flow, probe data will become a viable alternative to physical sensors over the next few years because the cost to deploy is much less while potential coverage is so much greater. But like I mentioned, we believe there's no "silver bullet" for traffic data; multiple data sources are needed for the complete traffic picture. Even perfect probe data isn't enough if you don't model the effect of that snowstorm arriving this afternoon.
GPSLodge - Are there any privacy concerns? Are you tracking me with my phone?
Connie Li - No, with both cell probes and GPS tracking, we have no knowledge of who it's tied to an individual phone or their personal habits. Besides, we're not interested in individual units as much as the modeling the mass of units moving down the highway.
GPSLodge - You mentioned that you have the ability to role Weather and Events into your Data Fusion Engine, how does that work?
Connie Li - To enhance forecasting capabilities, we've partnered with MyWeather and incorporated their MicroCast technology in to our Dynaflow solution. MicroCast enables road-specific weather forecasts, detailed to a latitude-longitude, and can pinpoint the likelihood of icy roads, travel-delaying snowstorms, dangerous fog and other weather conditions affecting travel. Dynaflow accurately estimates travel speeds based on the impacts of the range of dynamic factors impacting traffic such as incidents, weather and construction. We figure that weather accounts for over 25% of atypical driving times; even higher when you look only at predictive travel times.
GPSLodge - With all of this, why TrafficCast - how are you better?
Connie Li - We believe our foundation in traffic science is a real competitive advantage. More than data processing or statistical analysis, traffic science affords us perspective on how to make the best use of traffic impact data. We know what's available today, we can see where the data and applications are heading, and we're sensitive to bridging to the future so that our clients and their customers can have an increasingly satisfying experience that's based on traffic engineering fundamentals and not a promise of what's to come.
GPSLodge - Where is the best place to see TrafficCast in action?
Connie Li - We have customers in media, internet, mobile/telematics, public sector and fleet/logistics including Yahoo!, Google, Sigalert, Westwood One, MyWeather, Choice Logistics and GeoDecisions.
GPSLodge - Where does this all take us in ten years; what's the future hold?
Connie Li - Traffic information will become a standard feature of any desktop or mobile location based service. Analysts predict that GPS-enabled mobile phones will more than triple over the next five years and that LBS subscribers will rise from 16 million in 2007 to 43.2 million in 2008, and hit 300 million in 2011. We plan on continuing the charge to be the leading provider of traffic data on all those devices.
GPSLodge - Thanks Connie.
Read More in: Traffic News
Share this Article with others:
Came straight to this page? Visit GPS Lodge for all the latest news.
Posted by Scott Martin at July 14, 2008 8:30 AM