Predictive, streaming and in-memory analytics gain momentum

This past week brought an array of news in the Big Data and Analytics space, including new products, new funding and interesting new hires. The announcements weave a web between predictive, in-memory and streaming analytics, each of which is rapidly moving from cool technology to standard analytics tool.

Microsoft announced on Monday that it will be introducing a cloud-based machine learning service called – logically enough – Azure Machine Learning. Prepend the “Microsoft” moniker to the service’s name and its acronym becomes MAML, although the service doesn’t claim to be a sentient being. What it does claim to provide is a hosted service for predictive analytics, which gives Microsoft Azure a first-mover advantage over cloud rival Amazon Web Services.

I’m not certain, but this new service looks to be a clean break from the Data Mining component of SQL Server Analysis Services, a component which has been little-enhanced since 2005. Public preview access to MAML will open next month. I’ve signed up for this access and will report back once I’ve been able to get hands-on with the service.

Relevant recruiting
Speaking of predictive analytics, Seattle-based startup Context Relevant provides very impressive technology in that area. Rather than simply allowing users to build predictive models using algorithms of their choice, Context Relevant’s product analyzes data, and picks the best algorithm (and parameter settings) on its own.

Context Relevant’s solution is readily applied to algorithmic trading and risk management and, as such, has attracted the interest of many financial services firms. Perhaps it’s not surprising, then, that a few FinServ folks have seen fit to go work for the company.

Brad Spiers, formerly of Bank of America; Konstantin Getmanchuk, formerly of Goldman Sachs; and Bill Walrond, formerly of Deloitte Consulting, have joined Context Relevant as Vice President of Customer Success, Head of Financial Products and Engagement Director, respectively.

I think people in the database/analytics industry should keep an eye on this company. The tech is very advanced, and the leadership team is dripping with gravitas, especially now. 

Dollars and (Si)Sense
SiSense, an Israel- and New York-based BI and analytics firm with a product of the same name, picked up a cool $30M in Series C funding last week, led by DFJ Growth, with participation from existing investors Battery Ventures, Genesis Partners and Opus Capital. The product, which at first blush appears to offer only visual data discovery and dashboard capabilities, in fact embeds one of the more sophisticated single-node database engines available.

That engine implements not just flavor-of-the-month in-memory processing, but also utilizes a server processor’s on-board cache for even faster processing. Combined with its revamped front-end visualizations, this product, which has impressed me for years, is looking very strong right now. The market may agree, resulting in the very impressive growth SiSense has proclaimed, and the new funding it has received. (Full disclosure: SiSense retained the services of me and my former company, Blue Badge Insights, before I joined Gigaom.)

CodeFutures simplifies streaming data processing
Louisville, Colorado-based company CodeFutures has introduced its new AgilData product, that aims to make the whole premise of working with streaming data far simpler.

AgileData brings novel, powerful capabilities to streaming data processing, making it accessible to developers and even business analysts, rather than just Big Data specialists and Data Scientists. While lots of companies, including Amazon Web Services, are providing products and services for processing streaming data, they still require the customer to do the hard work of (1) connecting to the streaming data’s source and (2) processing the data after it has been captured. AgilData makes each of these tasks much easier.

AgileData uses an adapter-based approach to connecting to data sources, and it provides a SQL abstraction layer for the creation and querying of streams. Under the hood, AgilData uses the open source the Apache Storm streaming data processing project. Meanwhile, developers are shielded from that complexity and are able to work with data streams as if they were conventional SQL databases.

This is big stuff, though that magnitude of it may be lost on a market that is still drowning in streaming data buzzwords. Hopefully CodeFuture’s messaging can convey AgileData’s value in terms that are as simplified as the access it provides to streaming data in the first place.

Birst gets HANA-happy
Cloud BI provider Birst is announcing today a new version of its service designed specifically to work with SAP’s venerable in-memory database, HANA. Called simply Birst for SAP HANA, the company claims that it is the “first and only BI platform that automates the building of a data warehouse within HANA.”

My guess is that other BI players may take issue with that claim. Nonetheless, given HANA’s ability to run in the cloud and Birst’s native cloud orientation, customers of the two companies will now have the capability to do in-memory analytics on important ERP and other data, without requiring on-premises infrastructure or conventional BI software stack licenses.

New is the new old
Taken together, this week’s events show that the forward marches of cloud, in-memory, predictive and streaming data technologies are each gaining momentum. Things in these arenas are moving so quickly, in fact, that it begs the question of what big new thing may be next. The pace of change in 2014 is unprecedented.