Datameer 6 Delivers New UI, Spark Support

 
 
By Darryl K. Taft  |  Posted 2016-05-31 Print this article Print
 
 
 
 
 
 
 
big data BI

Datameer 6 provides a new user experience for iterative analytics and a re-architected, future-proof back end supporting Apache Spark.

With a goal of further democratizing big data analytics by making traditionally complex tasks easy, Datameer recently introduced Datameer 6 to enable a new class of data-driven business analysts.

The latest version of the Datameer big data analytics platform also features a new user interface (UI) on the front end and a future-proofed back end that now supports Apache Spark.

The enhanced UI delivers a new user experience that enables users to move fluidly between the various steps of working with data—from integration to preparation, analytics and visualization.

"On the user interface, we've totally modernized the look and feel of the product and done a lot to make the data the star of the show," said Andrew Brust, senior director of market strategy and intelligence at Datameer, in an interview with eWEEK. "Changes to font family, size, colors and layout mean that the data stands out and the rest of the user interface—including toolbar buttons, inspectors and dialogs—are unobtrusive and play their appropriate supporting roles."

Moreover, ease of use is becoming more important for business analysts who are overwhelmed with growing data complexity, said John L. Myers, senior analyst of business intelligence at Enterprise Management Associates.

"I was really impressed with the improved ability within the Datameer platform to allow someone doing data discovery to ‘move into and out of' data discovery components almost at will," Myers told eWEEK.

Having used multiple platforms for discovery and analytics, Myers lauded the Datameer 6 platform's ability to present data configuration information alongside of data profiling and analysis "tabs" without trying to remember where that configuration is. In many platforms, this represents a "restart" to the data discovery/analysis process. Datameer provides this visibility and configuration within a single view via multiple tabs, as opposed to attempting to find the beginning of a data flow and returning.

"If you have ever built a complex data discovery process, you know this is removes barriers to quick exploration adjustments and often progress toward your ultimate analytical goals," Myers said.

"For user experience (UX) we've added Context Tabs, which let the user easily shuttle back and forth between data management, prep, analysis, visualization and more," Brust said.

This UX reinforces the company's philosophy that data ingest, prep, transformation, analysis and visualization are iterative and reciprocal, rather than linear or even strictly cyclical, he noted. Users can—and should be able to—move laterally and iteratively in the order that most suits them and their train of thought.

"The flow is clearly more natural in Datameer 6," said Tony Baer, principal analyst at Ovum. "That said, I'd like to see more enhancements, like right-clicking to show the lineage of derived data sets."

Traditionally, the analytics process requires either specialized tools for each step that depend on different users and skill sets, or end-to-end-tools that require sequential workflows, Datameer said.

"It's not enough to just get data analytics in the hands of business users—it has to be extremely easy to use," said Stefan Groschupf, CEO of Datameer, in a statement. "The user experience has to engage them with their work and the goals of the company and the back end needs to handle all of the heavy lifting while removing the technical complexity."

Datameer 6 also introduced the addition of Spark to its patent-pending Smart Execution technology, which automatically selects the best processing framework for every single job while abstracting complexity from the end user, Brust said. This addition ensures the fastest processing time and allows the user to focus on the business rather than the underlying technology.

"This means that in addition to MapReduce, Apache Tez and our own single-node in-memory engine for small jobs, the analyses in Datameer workbooks can be executed on Spark, which works really well for a great number of workloads," Brust said. "Unlike most other products that support Spark, however, we have access to other engines as well and Smart Execution will use cost-based optimization to determine which engine or combination of engines will work best for a given job."

What this also means is that should some other execution framework come along and unseat Spark in the future, Datameer can integrate it into the Smart Execution platform and customers will be able to run their workbooks on it, including the workbooks they are building today, Brust added. So not only are they getting the power of Spark, but they also are getting execution optimization and future-proofing, freeing them to focus on their analytics work instead of deliberating which execution framework to run it on.

"The other part of the Datameer 6 features that I like is the architectural choice to include multiple processing engines in the Datameer processing optimizer," Myers said. "There is no one single processing engine that can meet all the requirements of analytical workloads. Using multiple engines on a single dataset without having an analyst move between applications or configurations saves time and continues to remove barriers to the data discovery/analytical process."

 
 
 
 
 
 
 
 
 
 
 
 
 

Submit a Comment

Loading Comments...
 
Manage your Newsletters: Login   Register My Newsletters























 
 
 
 
 
 
 
 
 
 
 
Rocket Fuel