Making Sense of IoT Data With Machine Learning Technologies

how the Internet of Things (IoT) will radically change your big data strategy?Massive amounts of data from sensors, wearable devices, and other technologies are creating new and exciting opportunities to make better business decisions in real time. However, harvesting all of this data is only half of the equation. Making the data actionable is where real value lies.

Traditionally, companies have mined data to look for trends and opportunities. In the world of IoT, searching for nuggets of information in petabyte sized databases is the equivalent of trying to find a needle in a haystack. To help extract value quickly and effectively, companies are turning to machine learning technologies, like big data technologies. However, implementing machine learning successfully can be extremely time consuming and complex. This has given birth to a new breed of vendors who deliver machine learning as a service, allowing customers to quickly implement technologies to turn massive IoT databases into actionable, revenue generating gold mines.

Machine learning as a service

Mining large data sets to discover business opportunities can be an expensive undertaking. Many companies who are starting to harvest large amounts of data from connected devices and sensors are not experts at big data management and machine learning. To get in the game they must hire data scientists to build a number of algorithms and various data mining capabilities with the hope of extracting value. As companies embark on this long journey, the valuable insights hidden in the data are driving up costs and not adding to the bottom line. How can these companies get these insights to market faster while reducing the risk of project failure?

One way is to leverage the expertise of companies whose core competency is machine learning. We live in a world where everything is being delivered as a service over the Internet. Companies have emerged that abstract away the complexities of executing machine learning technology. They provide easy to use services so that buyers can quickly get the insights they need without making huge investments in technologies that are not core to their business.

smarter

One interesting use case comes from Prelert, a self-described anomaly detection company. Prelert automates behavioral analytics allowing its customers to discover real-time insights while minimizing the upfront investment. Recently, Prelert teamed up with a major metropolitan city to help solve its traffic congestion issues and become a “Smart City” by studying patterns within its data. The city deployed various sensors to collect information about travel times of cars, buses, accidents, construction zones, and various other data points that impact congestion.

The city’s goal was to minimize congestion by prioritizing incidents so that higher impact events where addressed first, thus improving response times and reducing congestion. Prelert’s automated anomaly detection technology used machine learning and behavioral analysis to create complex statistical models of travel times while correlating traffic patterns in real-time with zero manual intervention or preset rules and assumptions. Here is a quote from Prelert’s blog that explains the process:

“The way this works in practice, for example, is on a given day at 6pm the Prelert engine identified and correlated significant increases in travel times on 9 of the more than 2,000 roads analyzed. It was then possible to prioritize and look at those 9 roads first, in order of severity, and tackle the congestion problems in a more helpful order.”

This short and target initiative resulted in massive reductions in traffic. Before machine learning as a service, the city would have had to hire a team of data scientists and invest large sums of capital to build the appropriate technologies in-house. Instead, the city leveraged Prelert, a company that delivers machine learning as a service as their core competency. The business problem was solved quickly, with minimal investment in time and resources.

Summary

Outsourcing non-core services is a reoccurring theme that I continue to highlight. We live in a world where the speed that business needs to move is much faster than the time it takes to conceive and launch new solutions in the areas of cloud, big data, and the IoT. The old build versus buy decision is transforming into “build versus rent” or “build versus managed services.” Add machine learning to the list of technologies you should consider outsourcing to the experts.