Hadoop meets Google Docs: Analytics made easy

Adatao's proprietary analytics suite gives Hadoop a Web-friendly front end

If Hadoop users have one persistent complaint, it's that the product is tough to work with no matter who you are, whether analyst, BI guru, or data scientist. Adatao, a company that recently landed $15 million in funding from various partners, including Andreessen Horowitz, is aspiring to make Hadoop as easy to work with as Google Docs.

The comparison with Google Docs isn't arbitrary. Christopher Nguyen, co-founder and CEO of Adatao, worked previously at Google as a director of engineering for Google Apps. With Adatao, he wants to offer an enterprise data-analysis product that makes it easier to extract useful results from Hadoop minus the tedium associated with using Hadoop -- though not at the expense of dumbing down the system to the point where the majority of Hadoop's functionality isn't available.

Nguyen explained his main motive was to "focus on the user first, and everything else follows," and to think of how to provide actual applications for Hadoop now that it has been established as a service. "Now that we have this massive query engine," he said, "let's make it boring" -- in the sense of being easy to use and predictable.

Adatao's products (which are proprietary, not open source) are not a single add-on to Hadoop. Rather, they constitute a set of layers, each designed to make Hadoop easier to work with in a different way. Queries can be written in a natural-language system devised by Adatao, but can also be executed via a number of other common languages used in data analysis: R, Python, Java, and SQL.

The top product, pInsights, is a Web interface that allows data to be queried from Hadoop and visualized. Resulting queries can be assembled into reports, charts, or other visualizations in much the same way one would assemble a spreadsheet or chart in Google Docs or Office 365.

The Web-office-suite metaphor is not intended to limit Adatao to the Web-driven presentation layer (à la Qlik Sense Desktop), but is meant to indicate real-time collaboration. Nguyen offered an example in his demo, where two people in remote locations collaboratively edited the same data visualizations as if they were working together on a spreadsheet.

Below pInsights sits pAnalytics, the second product and the one that performs the actual analytical work, which uses Apache Spark as its interface to Hadoop. pAnalytics can be accessed through its own set of APIs and be queried via all the languages used in pInsights. Adatao claims the big advantage of pAnalytics is in its presentation of the underlying data to the end-user as a simple table, so the user doesn't have to know how to run MapReduce or the other algorithms commonly associated with Hadoop data processing.

Nguyen noted that most of the company's customers demand on-premises deployments, in large part because of the general immobility of the data analyzed. "We talk to a lot of CIOs," he said, "and the responses has been for this kind of work that data has weight and can't be moved easily, or is sensitive competitive data. That sort of settles the argument for the initial go-to market to be enterprise or on-prem delivery, but there is a long tail of demand in the cloud as well."

Adatao isn't the only company trying to deliver solutions of this type. Platfora, for instance, has its own brand of self-service analytics. Likewise, Cloudera and Pivotal, two big-name Hadoop vendors, are calling attention to similar features in their own distributions. That's where Adatao is likely to face the most direct competition: not merely from other analytics suites, but from those gunning to deliver the most complete, polished, and end-to-end Hadoop solutions to enterprises as possible.

At the very least, investors are curious -- enough that a $13 million round of Series A funding in the company is being led by Andreessen Horowitz, along with Light Speed Ventures and Bloomberg Beta. The company states it's planning to use the cash "to grow [its] team and continue product development to further refine [its] offerings."

This story, "Hadoop meets Google Docs: Analytics made easy" was originally published by InfoWorld.

Join the discussion
Be the first to comment on this article. Our Commenting Policies
See more