In the coming year, the industry will more aggressively embrace IoT. “We’ll begin to see more drone and satellite data combined with other IoT and operational data to create more granular signatures and insights,” said Wael Elrifai, senior director of Pentaho Enterprise Solutions and AI at Hitachi Vantara. “This combination of big data and IoT will continue to unlock opportunities for companies to blend and analyze even greater datasets for additional insights using artificial intelligence—including machine learning and neural networks. For those that succeed, the next phase of IoT will be about moving beyond efficiency, productivity, and automation achievements to focus on bigger outcomes such as building more services and nurturing customer loyalty.”
Another aspect of IoT will be real-time response to IoT sensors. “The degree to which companies can ingest data from numerous IoT sensors can spell the difference between gaining good situational awareness and running blind,” said Augie Gonzalez, director of product marketing at DataCore Software. “Traditional serial batch processes that were once favored to centrally gather remote sensing data have proven utterly inadequate. Only through real-time parallel data access can one expect to keep up with high-fidelity inputs from edge devices, and react quickly enough to bypass malfunctions and thwart threats.” Gonzalez said there will be efforts to remove “long-standing choke points up and down the stack,” including employment of tools that “enable parallel I/O processing using standard multicore servers without having to change a line of code. This alone can measurably improve reaction time, even for legacy, locked-down systems.”
Location intelligence platforms
Among the next-generation technologies having the deepest impact in data-driven enterprises, location intelligence holds a great deal of potential for better understanding customer requirements, said Santiago Giraldo, director of product marketing for CARTO. “These platforms are built to analyze and visualize location data and allow business analysts to gain insights on where things happen, why they happen, and predict what will happen in the future.”
Self-service
The ability of end users to access and build their own interfaces—via self-service data preparation—is changing the enterprise analytics landscape. “Business analysts now have the power of an ETL tool on their desktop and can access structured and semi-structured data on-demand, including operational reports,” said Frank Moreno, VP of worldwide marketing, at Datawatch. An added bonus, he pointed out, is “the incorporation of gamification, which fosters a culture of participation and contribution.”
Real-time information
With the ability to capture and move large volumes of data from a variety of sources comes the need to view and analyze it in real time. There are a number of capabilities driving real-time analytics, starting with the growing number of “mobile devices such as tablets used on the plant floor, which have more real estate to show more KPIs and metrics and the detailed data table underneath,” said Elizabeth Vanture Cain, senior manager, product marketing, manufacturing, for Epicor Software. Add to the mix the proliferation of “enabling devices for real-time connected enterprises which include IoT sensors and live OPC connections to machines, and IoT hubs in the cloud as well as stream analytics and big data.” These can be supported with “analytics in the cloud to provide near real-time KPIs and metrics based on ongoing trends. These enable making decisions closer to real time, not months after the fact.”
Artificial intelligence
The year ahead will be a learning year for artificial intelligence and related initiatives. “The promise of AI, machine learning, and deep learning is undeniable but that promise won’t be fulfilled easily,” Joe Pasqua, executive VP of products for MarkLogic, cautioned. “There is rightfully a lot of focus right now on AI algorithms, but until you have the data, those algorithms aren’t very useful. We also have to understand that for the time being, organizations that wish to apply cutting-edge techniques are going to need experts to do the work. Right now, AI is more of a toolbox than a solution. We need systems that make it easier to gather, govern, cleanse, and correlate lots of different types of data from lots of different sources. Those systems should be more agile than the rows and columns of relational databases. And those systems need to be smart in the way they expose AI methods to their users so you don’t need a Ph.D. to use them.”