Big Data stalwart DataRPM Takes On IBM’s Watson With “Sherlock”
Much of the talk in machine learning and natural language software has been centered around the IBM supercomputer, Watson. But, DataRPM, a startup and self-described “industry pioneer in smart machine analytics for big data, thinks it’s time there was a Sherlock to take on Watson.
CEO and Co-Founder Sundeep Sanghavi were among the company representatives dressed in classic British detective costumes at this week’s Strata + Hadoop World Conference in New York, as part of DataRPM’s new “Sherlock” software launch campaign, which also included a clever Youtube ad spot and raffle trip to London–private Sherlock Holmes tour included, of course.
Sanhavi says that, like the iconic detective, the company’s cognitive algorithm backend allows users to “draw meaningful conclusions from their data quickly” by posing plain, natural language queries through a “Google-esque” interface. No programming or data science experience needed, he says, just formulate a straightforward question and receive an intuitive visualization in response.
“It’s elementary,” says Sanghavi.
The natural language component will likely be one of the stronger draws for business end users, but DataRPM also targets the needs of data specialists. According to DataRPM’s own estimates, modelling and integration alone typically consume 50-80% of the time invested in data analytics projects, especially as parameters shift and evolve.
DataRPM’s “automatic data modeling” software looks to reduce that load significantly. Cognitive analytic algorithms identify relationships across data sources–Hadoop, Google, Oracle, CSV, Salesforce, MySQL, etc–and and database tables, bringing together working models for a given query automatically, without the manual task of building the individual data warehouses. The models, stored as “computational search graphs,” can then be juxtaposed and cross-sourced in real time for single-page exploratory data analysis presentations highlight key measures, search, sort, and filter without any additional configuration.
The result is less HR demand overall and a data analytics team that’s free to focus on the real detective work of producing actionable insights for business operations.
Essentially, DataRPM is attempting to position itself as a smart software solution that integrates the business and pure data science elements of big data analytics. Just as data analysts are able to prioritize market insights over the grunt work of data aggregation, the natural language interface and automatic visualization tools make the eventual business intelligence accessible to the layman.
Cloud-based collaboration will play an integral part in fostering that dynamic. DataRPM has incorporated a drag-and-drop visualization dashboard into its software, complete with interactive filters and drill-down capabilities, for ad-hoc presentations and reports. Administrators can invite team members onto real-time discussion threads over a secure cloud network, to get feedback on individual data points and reports and clarify any confusions.
In any case, we are sure to see more of ‘Sherlock’ in the future.