Let’s face it: Visualization is a numbers game. Every analytics software company has its own way of presenting the numbers.
Datameer Inc., which makes a suite that pulls, processes and visualizes data from Hadoop stores, has a new way of visualizing data in the middle of processing: In can be flipped from spreadsheet view to a graphic view at the push of a button.
By seeing certain data graphically problems can be spotted quickly before business analysts get what they think are polished numbers, said Karen Hsu, the company’s senior director of product marketing.
The capability is included in version 4.0 of Datameer, released this week.
“What we’re trying to address in the 4.0 release is the challenge in workflow: People are spending up to 80 per cent of their time just doing data acquisition, management, data wrangling before they can visualize the data,” Hsu said in an interview.
“The challenge with the market today is there are multiple tools out there that address just a part of each workflow” — and most are designed for IT and not business analysts, and can’t visualize until that data processing has been done, which means she can’t see early that there might be dirty data.
For example, piles of customer data including name, address, age and sex looks innocuous. However, a small number of people might have filled out the column for “sex” wrong (or the machine reading software misread a form wrong.
Highlighting a column and pressing a Flip button makes the table flip as if it were showing the back of the page. But on the other side the number are in whatever graphic form the user wants — bar chart etc.
In that view it would be quickly obvious there are three bars for “sex,” so some data filtering has to be done before the final numbers are given to a business analyst to crunch.
If the data isn’t cleaned, the analyst may come to wrong conclusions.
Analysts can flip data after very integration, join and transformation. Other solutions only visualize data at the end of all transformations, Hsu said.
Through an optional Smart Analytics module for business analysts, users can also “flip” the spreadsheet to see an interactive visualization of the results from any advanced analytics, including decision trees, clustering, column dependencies and recommendations.
Datameer 4.0 will be available in Q2. It is sold in personal, workgroup and enterprise editions. It can be bought online or through Canadian partners including IBM and Dell.
