IBM Ups Cloud Data Portfolio for Developers, Data Scientists
IBM Predictive Analytics is a service that enables developers to easily self-build machine learning models from a broad library into applications to help deliver predictions for specific product use cases, without the help of a data scientist. "What we're driving at here is a way to get more folks to incorporate machine learning into their analytic operations and to enable those models to be deployed into applications," Kocoloski said. "So instead of asking someone to spin up a Hadoop cluster and then take the model they developed and figure out how to write an API to power it, we're leveraging some of the capabilities we have in our SPSS portfolio and taking that and ensuring that models that are built in SPSS can be exposed as APIs to applications. Also, with this, users who may not even be familiar with SPSS can start doing model development on the cloud to do signal extraction and work with data sets in a new way." He added that the Predictive Analytics service is one way that people can start to get more familiar with what it means to use machine learning algorithms and cognitive algorithms in the increasingly noisy world of data. Under the hood, it takes advantage of an auto-modeling capability in IBM's SPSS predictive analytics product. "We're trying to help people by proactively choosing an appropriate set of models that are most likely to be able to extract signals," Kocoloski said. "The bar we've set for ourselves is that it's not about the sheer scope of feature set and functionality here; it's rather about how consumable can we make this offering. We identified a new class of user and we want to get them more familiar with this world."The analytics exchange plays into the overall "open for data" theme and the desire to make sure that public datasets are immediately accessible and able to be incorporated into a broader analytics routine, Kocoloski said. "We think there is quite a lot of value to be gained from blending enterprise data with data from public and premium data sets," he said. "But often one of the challenges there is just knowing what data sets are out there. So with the exchange we've built a comprehensive catalog of datasets, and it also sets the foundation for better cataloging and metadata management in the cloud native fashion. That's an area that's ripe for innovation." The new offerings build on IBM's investment in Apache Spark and further complement its mission to provide enterprise-class support for open-source developers and data handlers of any level. IBM has redesigned more than 25 of the company's core analytics and commerce solutions with Apache Spark—helping to dramatically accelerate their real-time processing capabilities. "Data is the common thread within the enterprise, regardless of where its source might be. In the past, data handlers have relied on disparate systems for data needs, but our goal is to move data into the future by providing a one-stop shop to access, build, develop and explore data," said Derek Schoettle, general manager of IBM's Analytics Platform and Cloud Data Services, in a statement. "IBM's integrated Cloud Data Services give developers greater scalability and flexibility to build, deploy and manage web and mobile cloud applications, and enable data scientists to apply information across businesses efficiently."
Finally, the IBM Analytics Exchange is an open data exchange that includes a catalog of more than 150 publicly available datasets that can be used for analysis or integrated into applications.