Time to Talk: Tableau Prep
For the last few months we have been waiting patiently with great anticipation for the release of a new Data Prep tool from Tableau, originally called Project Maestro, Tableau Prep has now been released, and in some ways, it's been worth the wait.
Interface and Ease of Use
The first thing that we like is the interface design, it is in many ways the data source window from Tableau Desktop but enhanced and expanded, so if you are familiar with Tableau Desktop you will find the interface easy to get to grips with. The main Part of the window is the flow designer this is where we create the discrete steps of our data prep process.
Down the bottom of the screen is a much expanded version of the preview window, with a look at the data outputted from a step at the bottom of the screen, just above this is what I’ve been thinking of as the data quality metadata window, where we can see all the unique values within a column and how often they occur.
This metadata isn't just informative, you can use it as a tool for data prep.
Right-clicking on an item in the window we can filter out or keep only a value, or even directly edit a value, you can also see a little bit more data on the rows that have the particular (in this case null) value in other columns.
The next thing we like is the simplicity of adding new steps, to do so you simply click the plus sign after a step.
There are several distinct categories of activity,
An interesting thing to note is the ability to do some aggregate calculations here and so expands the use cases for doing analysis in this step before the full Tableau experience.
Which brings us to the next thing we like about Tableau Prep
Calculations and Filters
The calculation window remains unchanged from Tableau Desktop, though it is worth noting that table calculations and Level of Detail calculations are not available to use in Tableau Prep, and the calculated fields here cant use aggregation functions.
Really this limits the calculations in Tableau Prep to row-level calculations to perform some basic data prep activities like concatenations truncations and date manipulations. Splitting of Fields is also available here.
Filters also operate slightly differently in Tableau Prep, being more formula-based, and users are encouraged to create boolean formulas to describe the filter they wish to achieve:
Drag and Drop
Much like desktop prep allows for dragging and dropping into unions and joins, simply grab any step or input and drag onto a union/join to add onto that step, or you can drag and drop any step onto another step to create a new join or union:
Related to the above, much like desktop undo works nice and simply and without fanfare.
Preview in Desktop
The output from any step can be previewed in Tableau Desktop by simply right clicking on the step.
Doing this creates an anonymous .hyper output and opens Desktop with a connection to that data source.
There are several options for outputting either to file or to a server, for outputting to file you can export to both hyper and tde. You can also export to csv.
See the Steps that have been done
Finally, the thing we like the most is the way you can interrogate what activities have taken place in each step, firstly you can see from the icons above each step the task types that have taken place, and hovering over will show a quick summary of what has been done:
For more detail you can see all the steps that are taking place in sequence for that step:
So while there's a lot to like and we are looking forward to using Tableau Prep for a wide variety of use cases. There are some areas where Tableau Prep doesn't quite hit the mark:
It would be good to include some ability to create table calculations and LoD calculations within this tool, as it stands now to add these calculations users are required to use Tableau Desktop before uploading the data source to Tableau server.
A few more tools for data preparation, including the ability to remove spaces from within fields (useful for handling phone numbers), and potentially the ability to add a record ID (a useful tool for more heavy duty data prep)
So overall a fantastically intuitive tool for handling some of the more common data prep use cases we are really excited to see how Tableau Prep is going to evolve over the next few releases, expect to see more tutorials on Tableau Prep in the very near future.