Using Azure, Microsoft and partners work to make traffic jams a predictable affair

Microsoft is in the midst of rebranding itself. The company wants Microsoft, the brand, to stand for more than just Windows and Office. As the world moves towards a combination of mobility devices and cloud-connected services, Microsoft is also positioning itself to become synonymous with ‘the cloud’ and big data.

Over the past couple of years, Microsoft has been partnering with governments, healthcare, education, aviation, sports and analytics firms to help provide the connected cloud platform for researchers and businesses alike, to work on big data. Recently, Microsoft Research partnered with the Federal University of Minas Gerais research in Brazil, to continue that growth. The two research firms came together to tackle an everyday occurrence in most places; traffic jams. In a study done by the Brazilian National Association of Public Transport, researchers found that traffic in the country accounted for a $7.2 billion (US) economic toll back in 1998. Since then, the traffic situation in the country has only gotten worse according to Bazilian traffic expert Fernando de Oliveria Pessoa.

Microsoft and the Federal University of Minas Gerais research are looking to combine traffic data from all possible avenues. Once collected, the data should help create a predictive program that can assess traffic conditions 15 minutes, to an hour out, for potential drivers. Data points will be gathered from transportation departments, historical and current information, as well as Bing traffic maps (not sure how well they work in Brazil), road and camera sensors and lastly from social sharing sites like Waze and others.

University of Minas Gerais research is taking terabytes of data and using Microsoft’s Azure scalability, immense storage capacity and computational power to help make sense of traffic in certain areas of Brazil. However, since Microsoft Azure is a cloud-based platform, real-time information and traffic predictions can potentially be shared with users, anywhere in the country or around the world.

The research model being used by the two research partners, has garnered an 80 percent accuracy in predictions. The predictions are not only limited to Brazilian traffic, but also other heavily congested areas in Los Angeles, London, Chicago and New York. As of right now, those results do not factor in social sharing sites. Researchers are encouraged that they can obtain a 90 percent accuracy rate once the social aspect of shared information has been included.

For some of us lucky Cortana users, we see some of this research currently being applied. Perhaps with a better accuracy rating, Cortana will be able to plot alternate routes and public transportation suggestions on the fly, to help navigate drivers through traffic riddled areas.