tableau-10-thumb.jpg
 Image: Tableau
PROS

  • Cleaner, leaner, user interface
  • Tableau Mobile for Android and new device-specific designer tool
  • New and improved data connectors
  • Cross-database joins
  • Filtering across multiple data sources
  • Data clustering

CONS

  • Still limited when it comes to predictive analytics
  • Browser interface still lacks functionality

Price

From $999 per user for the Personal Edition of Tableau Desktop and $1,999 for the Professional Edition


In the face of growing competition, Tableau has taken action to maintain its market-leading position by releasing a comprehensive upgrade to its highly popular data discovery and visualisation product. Among the more obvious changes are a revitalised user interface and improved support for mobile devices, but there's a lot more besides going on in Tableau 10 with the vendor both tweaking and adding to its feature list, mostly in response to feedback from a very active user community.

tableau-10-main.jpg
Tableau 10 Desktop
 Image: Alan Stevens/Tech Pro Research

Looking good

Given that Tableau prides itself on the usability of its visualisation tool, a radical re-working of the Tableau Desktop GUI might seem a retrograde step. Despite claims to have "reimagined the entire interface", however, the changes are subtle and most of the tools stay where they were, with some -- like the revised Show Me toolbox -- getting out of the way until needed. There are changes also to fonts, with Benton Sans used throughout the interface and a totally new font (unsurprisingly called Tableau) commissioned expressly for use in visualisations in this release.

tableau-10-old-new-gui.jpg
Here's the revamped Tableau 10 interface with the old GUI for comparison behind it. Check out the GUI changes for yourself using a comparison widget on the Tableau website.
 Image: Tableau

Other improvements include new iconography and a totally revised colour palette, with the end result being a much cleaner and more modern-looking interface that existing users should adapt to without too much trouble.

SEE: Big data policy

Long-overdue tools to apply font style and colour changes to all the sheets in a workbook also make it into this release, plus there's an all-new Device Designer to see how dashboards will be displayed across devices and to create custom layouts for optimising visualisations. The same dashboard is employed, accessed by a single URL, with Tableau 10 automatically applying the correct layout to match the hardware involved.

tableau-10-device-designer.jpg
With Tableau 10 you can check how dashboards will be displayed on different mobile devices and customise the layout to suit your needs.
 Image: Alan Stevens/Tech Pro Research

On the subject of mobiles, the Tableau Mobile app is now available for Android (phones and tablets) as well as Apple iOS. The apps can be downloaded for free, but both mobile and browser-based access require either Tableau Server (from $10,000 for a 10-user license) or a subscription to Tableau Online (the hosted version of Tableau Server) at $500/year per user.

tableau-10-mobile.jpg
Tableau mobile is now available for Android devices to connect to Tableau Online or an in-house Tableau Server.
 Image: Alan Stevens/Tech Pro Research

The server also benefits from the Tableau 10 update with much-needed stability enhancements, plus quicker upgrades without the need for lengthy backups and manual removal of the existing product when updating the software. New desktop licensing tools have also been added and there's a simplified procedure when publishing workbooks. However, existing 32-bit servers must be upgraded to 64-bit to work with Tableau 10.

Tableau is also working to replicate its desktop interface in a browser, with several additional features added in this release. As well as an updated look and feel, for example, web users can now author, save and share dashboards, as well edit existing dashboards from their browsers. Published data sources can be added and changes to previously saved workbook versions rolled back, but Tableau Desktop is still needed to take advantage of all the tools the product offers.

SEE: Quick glossary: Big data

More about the data

Already well featured when it comes to support for data sources, Tableau 10 adds even more starting with connectors for Google Sheets and QuickBooks online. Specific named connectors for fast data sources MemSQL and Presto have also been added to the list, along with another for specialist marketing platform Marketo. Some existing connectors are also updated, including SQL Server, which gets support for contained databases, and Oracle where table functions can be used to simplify access to complex data sources. SAP Hana support for Mac is another new feature, and Tableau also now stores the connection details for each data source to save having to type them again each time a connection is made -- a small but very welcome time-saver.

tableau-10-connectors.jpg
Connectors for Google Sheets, QuickBooks Online, MemSQL and Presto are all added.
 Image: Alan Stevens/Tech Pro Research

With so many data sources to choose from, it's inevitable that customers will want to apply analytics across more than one -- something that's already possible from within Tableau using the blending tool, but far from straightforward and limited in terms of functionality. As of this release, however, so-called 'cross-database joins' can now be created, allowing two or more data sources to be logically linked simply by adding them to a workbook.

tableau-10-cross-database-joins.jpg
Cross-database joins allow users to work with multiple data sources as if they were one big database.
 Image: Alan Stevens/Tech Pro Research

One of the most requested features by the Tableau community, cross-database joins make light work of this otherwise complex task without the need for external data warehousing or cumbersome data blends. Instead, it can all be done from the desktop with Tableau colour-coding the columns from each new source added to a workbook and displaying them side by side in the data grid. The resultant mashup can then be treated as though it were a single database, enabling users, for example, to perform row-level calculations and create visualisations using fields from any of the joined sources. You can even create an extract of a multi-connection data source, publish it to Tableau Server or Tableau Online and schedule it to be refreshed as though it were a single database.

On the downside, there are restrictions on which data sources can be included in cross-database joins: those held on Tableau Server, for example, are currently excluded along with Google Analytics, Salesforce and a few others. Performance could also be an issue, especially where large databases are concerned, as joining is done locally.

Still, there's no denying that cross-database joins are a great new feature and sure to come in handy. As will another -- the ability to apply filters across multiple data sources. The databases involved need to share a common dimension, but whereas a separate filter was needed for each source in previous versions, just one is required in Tableau 10, with the software also able to apply filter edits to all worksheets employing the same filter.

SEE: Job description: Data scientist

Better analytics

Last, but by no means least, Tableau 10 delivers numerous enhancements when it comes to discovering and visualising trends within data. Heading the list here is a new clustering feature able to automatically find and group together similar data points.

For the data scientists out there, this uses k-means clustering with a variance-based partitioning method for consistency. The rest of us, however, don't need to know that because the analytics involved can be applied simply by dragging the new clustering tool from the analytics pane and dropping it onto a visualisation, which can be of any type.

tableau-10-clustering.jpg
A new clustering feature makes it easy to find and highlight groups of similar data points.
 Image: Alan Stevens/Tech Pro Research

Elsewhere, bar charts have been made more intelligent in Tableau 10 and there's a new highlighter tool to spotlight specific data without removing other detail from a visualisation. Maps are enhanced too, with additional postal codes and more detailed geographic data. Tableau 10 can also match column names to geographic roles using languages other than English, while users can define their own geographic areas to, for example, configure multi-national groupings, sales territories and so on.

There's a lot more besides, and more on its way -- updates to the original 10.0 release are already being delivered to help keep Tableau ahead of the pack. The user community appears to love it, but the product does have its weak points, particularly when it comes to embedded analytics (although recently announced new APIs should help here). Beyond that, Tableau 10 does little to address the lack of built-in predictive analytics, but more advanced tools like clustering are a step in the right direction towards addressing this. Overall, Tableau 10 looks set to be a hit.

Read more