Tableau Aims to Speed Analytics with V9
Tableau Software has long positioned itself as a provider of tools that help you visualize big data, but hooking up to big data sources has not always been easy, and the resulting analyses have not always run as fast as people would like them. With the upcoming release of Tableau 9.0, the speed of analysis will be a big point of emphasis.
If you had to pick one tool that all big data users seem to have, it’s Tableau’s. With more than 26,000 customers around the world, Tableau’s brand of data analytics and visualization clearly resonates with data analysts, including a good number who use it with Hadoop. Indeed, the number one driver of new workloads to Cloudera is customers who want to run Tableau on top of CDH, a Cloudera VP recently said.
But as popular as it is, there are stories of people getting frustrated with slow response times on Tableau queries, especially when it comes to Hadoop. In many cases, the slowness has nothing to do with Tableau—a malformed SQL job running on Hive is often the answer. But in any event, performance is always important, and it’s one of the focuses of Tableau 9.
With Tableau 9, the company says it has completely architected Tableau Server and Tableau Online to deliver “a faster, more scalable” platform. The company says that customers will notice “a significant speed improvement,” which are due in part to the use of parallel queries to use available multi-core processors; a technology called “query fusion” that lets multiple queries at the same level of detail be grouped together and executed together; the use of vectorization; and smarter query caches.
How much faster is Tableau 9? Well, that will vary, of course, but some sources have Tableau 9 running three to four times faster than previous 8.x releases; the query fusion technology reportedly delivers a 50 percent boost by itself. The director of data analytics at software training provider Pluralsight says the same workbooks in version 9 take seconds to load, compared to the minutes it took version 8.x to load. “Tableau 9 is insanely fast,” he said. “This kind of blazing performance is exactly what we need.”
Other proposing new features that will debut in version 9 include a new Data Interpreter function, which automatically clean up dirty data for the analyst. Tableau says Data Interpreter, together with the Automatic Data Modeling feature that debuted with version 8.2, will help users do less data prep and more actual analysis of data.
Tableau is also highlighting a few other features that will keep analyst “in the flow” instead of requiring them to stop what they’re doing to perform some action. Chiefly, Tableau is allowing users to type their calculations directly into the screen and have their data update automatically. The company has also introduced a new “analytics pane” to make it easy to drag reference lines, totals, and other analytical objects right onto the screen.
Connecting to new data sources should be faster with version 9 thanks to the “add new data source” option Tableau added. And data scientists will appreciate the new connectors that Tableau is making available to SAS, SPSS, and R environments. For a good compilation of some of Tableau 9’s other functions, check out this story in The Information Lab.
Tableau is expected to announce Tableau version 9 tomorrow; the software should be available by the end of the month.