There is a lot to be excited about in the latest Tableau release. The list of new features is long and varied, and as a beta-tester I am studiously testing them one by one. Naturally, I am heavily biased by my own experiences and the specific needs of my clients -- a few of these features really stand out. Here are five (in no particular order) that have me anxiously anticipating Tableau Version 9.
#1: Level of Detail Calculations
There is a lot of new functionality for calculations including freeform input, interactive editing, and a new calculation editor, but far and away the best thing coming is the ability to define calculations at specific levels of detail.
In Version 9 you can specify how to aggregate at fixed dimension levels, including or excluding specific values. This will eliminate a lot of complex table calculations, scripting, and set building. With previous versions of Tableau, I had to repeatedly bring in the same data source in a single Worksheet to provide different levels of detail for calculations. This can slow performance considerably. LOD calculations eliminate a lot of this needless complexity.
This new functionality may be threatening to some Tableau Jedi who make their living by grokking partitioning and addressing schemes for complex calculations. However, the more I dug in the more I realized that this is a solid win-win for everyone who uses Tableau. Tableau developers can now spend their precious time on even more impressive visualization techniques and data analysis. This is a game changer.
Here is a simple example of how to calculate the sum of Sales by Quarter without worrying about how your viz is partitioned, addressed, or aggregated:
#2: Improved Data Prep for Excel
Most of my work with Tableau utilizes large corporate analytical data structures in major database systems like Oracle, Microsoft SQL Server, Postgres, or Hadoop, yet I can never escape Excel! It permeates every corporate environment I work in. No matter how wonderfully designed the Data Mart, there is always some kind of analysis in Excel that needs to be integrated somehow.
Tableau makes this incredibly easy with Data Blending, but prepping the Excel files for use in Tableau was always a complete pain in the you-know-what. I am constantly running Excel worksheets through Pentaho PDI transformations in order to clean them up for use, and I even installed Alteryx for the explicit purpose of unpivoting Excel for Tableau analysis for one client.
Talk about tool overkill, it was like killing ants with a sledgehammer. I love Alteryx and PDI, but there are many better things to justify their use than prepping Excel files.
In Version 9, I will be able to use Tableau’s new Data Interpreter to do my Excel data prepping without having to reach for my heavy-hitting and often expensive alternate tool sets. The Data Interpreter is a new option made available when you select your Data Source.
The Interpreter automatically removes blank rows and extraneous header rows from Excel files. This eliminates a lot of time-consuming manual manipulation. Another amazing new feature is the ‘Pivot’ option, which allows me to unpivot data into a much more digestible and analyzable format. Check out this example where I unpivot Quarterly Sales data:
See how that worked? I still had to rename the fields and I will perhaps manipulate the Quarter column to make it return ‘Q1’ or ‘1’ instead of ‘Quarter 1 Sales’, but this is going to save me a lot of time.
#3: Split Fields
In the example above I have a field that returns ‘Quarter 1 Sales’ as a value. I would rather have a header of ‘Quarter’ and a value of simply 1, 2, 3, or 4 without those extra words Quarter and Sales. If I highlight the Dimension ‘Quarter’, go to the new ‘Transform’ option and select ‘Split’, Tableau will automatically create a new Dimension field for me named ‘Quarter 1 – Split’ that contains only the quarter number I desire.
This can also be used to split full names or any string that needs to be separated. This may seem trivial, but I have lost count of how many times I have manipulated the underlying data in Pentaho PDI in order to achieve this result for various dimensions and zip codes.
#4: REGEX functions
OK, admittedly, this one is a little geeky. For anyone who has struggled with string pattern matching and less-than-optimal data sources, Tableau 9 has 4 new REGEX functions that you can use in calculated fields:
REGEXP_EXTRACT_NTH(string, pattern, index)
REGEXP_REPLACE(string, pattern, replacement)
As with the Excel Data Prep and the Split Fields, this was something I often had to accomplish in Pentaho PDI prior to loading the data into Tableau, regardless of data source. Hooray for one less data manipulation step!
#5: Query Performance Enhancements
Tableau has been touting multiple underlying performance improvements for Version 9 including Data Engine Improvements, Parallel Queries, Query Fusion, and External Query Caching. Naturally all of these improvements are dependent on data sources and will vary considerably depending on the nature of your underlying data is and where it is located.
This made me very cautious of Tableau’s claims of 10x or 9x or 5x improvements on very specific setups. Those claims make me want to test on larger and more diverse data sets than I have been using up to this point. I believe such testing will warrant its own blog post when I get some hard findings.
Anecdotally based on the few data connections and data sources I have migrated from 8 to 9 I am seeing noticeable performance improvements on my dashboards. This does not appear to be Tableau blowing smoke. It looks like Version 9 is faster in general. I like faster.
I could go on and on about the new Map Search capability, the new and improved security permissions, the greater transparency of the underlying PostgreSQL data dictionary, and a host of other items that are pretty exciting. But, all said, these five improvements are the ones I am most looking forward to in Tableau Version 9.