It’s important for data scientists to work together with IT departments and engineers to extract the most value from data gathered by IoT deployments, according to IDC’s 2019 predictions about the Internet of Things (IoT).
The data-processing aspect of the IoT, is going to be the central pillar that makes IoT worthwhile for businesses, said IDC group vice president of IoT and Mobility Carrie MacGillivray during a webinar Tuesday, and processing that data in a meaningful way requires the use of machine learning and artificial intelligence.
The problem is there aren’t enough skilled professionals to make every AI/IoT implementation work, according to MacGillivray, so businesses generally adopt one of three options: Putting existing, on-staff data scientists to work, outsourcing ML model-building to a professional services team or experimenting directly with open-source ML models.
But more and more, a range of engineers – mechanical, electrical, software, systems – are coming out of universities with AI and ML skills, so IoT analytics management is likely to shift to engineering teams, she said.
IDC expects businesses to take this fact on board quickly, and by 2020, companies are predicted to reach a 90 percent success rate implementing AI-enabled IoT systems. The current disconnect between the future of IoT analysis and the future is largely one of emphasis, MacGillivray said. The focus, particularly in sectors like manufacturing and fleet management, has been on getting every machine or vehicle connected as quickly as possible and worrying about getting detailed information out of the system later. Hence, one of the most common early applications of IoT tech has been on predictive-maintenance analytics.
But there are many more diverse applications of IoT in the