After about two months in public beta, Tableau just launched its 9.0 version officially. This release cycle has been focused on polishing the user experience rather than introducing a large amount of big new features, though there are a few significant new ones alongside UI and performance improvements. This is a good thing, as speed and ease of use are critical to let analysts focus on what the data has to say rather than fight with their tools.
The additional control over how story points look may first appear like the most visible change from the perspective of using Tableau to create narratives for publishers and their readers. Having control over the size, fonts and colors in story points is nice and will make embedded Tableau stories easier to look good within your site design. Here’s a quick illustration below, though arguably this should have been part of Story Points right when they were added in Tableau 8.2:
We’re much more excited about what’s going on under the hood, starting with level of detail expressions whose purpose is to allow the blending of broader and more granular aggregates in one chart, independently of the level of aggregation set up for the visualization as a whole. These “LOD calculations” remind us of how Power Pivot’s Data Analysis Expressions (DAX) can be applied to handle different granularities, or QlikView’s aggregate function.
This concept may sound overwhelming for those not familiar with it. In a nutshell this is powerful stuff that you cannot do in a regular spreadsheet, with which one can dig up insights otherwise buried in flat data. Mixing levels of details comes handy to find the reasons why aggregate data changes over time. It’s good to show how data evolves, it’s better to explain the drivers behind these changes. Of course such advanced features are relevant mostly given more detailed, bigger datasets spanning longer time frames. Those will typically turn into content restricted to your paid customers rather than free articles. Here’s a video if you’re interested in the implementation nitty gritty.
Much faster data tooltips and instant average/median calculations applied to the current selection are two very welcome speed improvements to help drill down within a broader data set. The latter feature is similar to Excel’s automatic tallies at the bottom of the screen, based on cells selected by the user. A new inline calculations editor also looks pretty close to the Excel functions editor, what with its color coding and auto-complete. But it’s pretty cool to see it applied to a whole visualization, as the animated image below illustrates:
We haven’t used the data preparation feature yet, which promises to replace third-party tools (including Tableau’s own Excel reshaper) for the often necessary process of cleaning up messy source data. It’s a good idea to do everything in one place, though I suspect many Tableau users are like us fairly proficient with doing all that prep work elsewhere, whether in Excel, Open Refine, or more sophisticated tools such as Alteryx. Here’s a video demonstrating this functionality, which includes impressive pivoting and splitting capabilities. Regular expressions can likewise help break down poorly-formatted data into chunks you can use.
As you’ll see in our entry explaining the whole life cycle from raw data to advanced visualization, data preparation, from merging to cleaning to categorizing data points, is crucial to proper analysis and charting.
We Like What We See!
Overall, our initial impression is that this is a solid release built with a good sense of priorities, not least because it’s faster in many ways. Unfortunately we are still facing a bug, already present with Tableau 8.3, where we cannot edit or move annotations once they have been created. We can always add static annotations in an image editing software when we create PNG charts from Tableau visualizations, but of course native annotations that work properly would be much preferable. This will hopefully be sorted out soon, and overall Tableau continues to improve to serve data journalism needs close to our heart, whereas our use case barely feels like it’s even on the radar of most of its competitors. Microsoft has got their work cut out for them with their Power BI beta!
Here’s the full list of improvements and additions, including a host of mapping improvements that we haven’t covered above. Some of these improvements affect what visualizations will look like to end users/readers, but productivity improvements behind the scenes are even more important to be able to do more than publish visualizations in a very ad hoc way that turns out to be very time consuming and expensive. At Needle Stacker we’re always looking for ways to “productivize” our services so that data can be fully put to work for our customers while remaining within the constraints of their business model.