Using Big Data and Cloud Computing to Predict Traffic Jams

Microsoft Research is working with Brazil's Federal University of Minas Gerais to tackle the seemingly uncontrollable problem of traffic jams. The Traffic Prediction Project plans to leverage all available traffic data - including both historic and current information gleaned from transportation departments, Bing traffic maps, road cameras and sensors, and the social networks of the drivers themselves. The immediate objective of the research is to predict traffic conditions over the next 15 minutes to an hour. By using algorithms to process all these data, the project team intends to predict traffic jams accurately so that drivers can make real-time choices. Achieving reliable predictions will involve processing terabytes of data, and the researchers are using the Microsoft Azure cloud computing platform for the service. To date, the researchers have tested their prediction model in New York, Los Angeles, London, and Chicago. The model achieved a prediction accuracy of 80 percent, and that was based on using only traffic-flow data. The researchers expect the accuracy to increase to 90 percent when traffic incidents and data from social networks are folded in.

Transcript

00:00:07 [Music] traffic jam is a common problem for big cities not only in Brazil but in the world people get stuck in traffic them all the time and they lose a lot of time we have economic losses and and pollution tackling this problem is a challenge every day you have more cars in the roads and the streets and it's really

00:01:12 hard to change the infrastructure to support that so right now we are jointly studying this problem the main point you are trying to answer here is how we can predict the traffic condition in this the near future to predict the future basically you have to know the the [Music] best the traffic prediction project uses

00:01:44 data available by social networks Department of Transportation and data that the users create themselves while they move around the city the idea is to combine all that and to create a solution in real time helping the users to get from point A to point B and a more efficient way this project uses Microsoft Azure as the platform to enable it as a

00:02:08 service Microsoft Azure helps us a lot in building the traffic prediction model because it is a lot of data for instance the traffic flow from Big Maps traffic incidents information available through sensors on the roads social networks for instance for square Instagram where people can share their locations all those kind of informations can be combined to forecast the traffic jam for

00:02:35 the next 15 minutes 30 minutes 1 hour we need to collect data from long periods of time and process them given the processing and storage capabilities of azure this was fundamental for our project the main series that I applied the prediction models are New York Los Angeles London and Chicago the model presented great results more than 80% of accuracy using only traffic flow data

00:03:08 and we still are working on the characterization of other sources such as traffic incidents and social networks for