How big data can end gridlock on the roads

Traffic is the bane of modern life, which makes Inrix’s slogan pretty attractive. The firm wants to “solve traffic worldwide”. It is doing so using an extraordinary mix of data sources to model traffic flow and ensure that every driver can be given a tailor-made route so they reach their destination with the minimum possible disruption. Founded in 2004, Inrix is the sole supplier of traffic data to the UK Highways Agency, and provides services to satnav maker Garmin, the BBC and ITV, and over half the state departments of transport in the United States. Hot sector? Just consider that Google recently paid $1bn for Inrix’s much smaller rival Waze.

The traditional approach to monitoring traffic involved using road sensors, which are expensive and barely worth installing on minor roads. Inrix uses GPS signals, which each car with a satnav, fleet-tracking device or configured smartphone app provides. Inrix sucks in real-time GPS data from 100m vehicles. This gives a picture of traffic flow in the UK, the US and other nations to within 15 seconds’ accuracy. And there’s more. “We have over 100 sources of data creating trillions of data points”, says Inrix’s general manager of big data Kevin Foreman. These include weather reports, Twitter, event listings, emergency services data, construction works, commercial fleet activity and other sensors.

By using non-traffic big data sources it is possible to increase accuracy. “We do forecasts just like the weather guys do”, says Foreman. “We want to be able to tell you what a road will be like in three months’ time at 5pm on a Tuesday.” Take events. “We use event listings to factor in things like Manchester United playing a game on a Wednesday evening. When they play, we can alter our model to predict what the impact will be and advise drivers.” Twitter is used too, to identify flashpoints. If the M25 is being tweeted a lot, Inrix will try and identify why.

Getting from A to B faster is the obvious benefit. But traffic data has many other applications. Take billboards. Advertisers want to know how fast cars are going past a billboard. Board owners could charge per car. Or delivery firms such as FedEx and DHL, which currently charge a flat rate per job, can alter prices to reflect fuel consumption and staff costs based on actual road conditions. Even the real estate industry is asking for the data: buyers can know not just how far they are from work, but how long in terms of driving time.

“We are getting real interest from the car makers”, says Foreman. “Cars monitor their brakes, ABS, windshield wipers and fuel consumption. Imagine if we could combine that data with ours. Suppose it snowed. Currently, the authorities clear ice according to a rigid pattern. But with this data they would know where cars are using ABS a lot and travelling slowly. They could clear roads in order of severity by using real-time data sent from the cars themselves. Or fleet owners could identify drivers who are going too fast, or using their brakes too hard, and advise them before an accident occurs.”

Solving traffic is just the tip of the iceberg.

Charles Orton-Jones is a business journalist.

Image source.

Share this story