Tableau recently released a new feature "Quick LOD" that should make it easier to build LOD calculations. As an intensive LOD user, I want to try this feature out and share my thoughts with you.
Content:
· Quick reminder about LOD expressions
· Testing "Quick LOD"
· Testing Outcomes
· In a nutshell...
Quick reminder about LOD expressions
When we build a data visualization we often need to dispose of multiple aggregation levels of the same measure at the same time, by using LOD calculations.
If we make a conceptual parallel with SQL, LOD expressions allow defining more than one different GROUP BY clauses in the same query. Or, using more advanced SQL concepts, LOD would allow multiple Windows Functions (with some differences thought) on the same data set.
For instance let’s consider a viz about accommodations, we have a data viz about the number of reviews per room type and neighbourhood. Here I use a ugly but clear representation:

If in the same data viz we want to display also the aggregation by neighbourhood only, we would need a LOD expression.

OK, it doesn’t look very useful at first sight, but we could use that "LOD Number of Reviews" for other more complex and interesting calculations (e.g. weighted averages by Neighbourhood, % out of totals, etc.).
There are 3 types of LOD calculation, to specify how the data group should be built:
- "FIXED" LOD defines a static and independent grouping. It computes the aggregated measure by using only the dimensions specified in the calculation, without reference to any other dimensions in the data visualization. Also, FIXED expressions ignore all the filters in the viz, except for context filters, data source filters, and extract filters.

- "INCLUDE" LOD groups data to reach a more granular level. It computes the aggregated measure by using the dimensions specified in the calculation, in addition to whatever dimensions are in the view. So it’s useful when you need a calculation that includes a dimension that isn’t in the view.

- "EXCLUDE" LOD groups data to reach a less granular level. It explicitly removes dimensions from the expression, meaning that the specified dimensions are excluded from the grouping.

In-depth explication of LOD are outside this article’s scope, but if you need more info, here you go:
Testing "Quick LOD"
I have been using Tableau for 5 years now, and in terms of expression complexity and possibilities offered, LOD were the concept it took me longer to understand and use fluently. Now, as I often deliver trainings that also include Tableau, I tried the new Quick LOD feature to see if it can really help beginners to use LOD, or at least to get started.
What Quick LOD does is that it simply creates a new measure with a standard LOD expression by drag and drop a measure onto a dimension. From the tests I did, the default LOD type selected is "FIXED".
If you’d like to test them:
- Open Tableau Public and download the Sample Data Set "Airbnb Listings in New York City".
Note: in case you are a complete beginner, I described how to get Tableau Public (free) here.

- In a new worksheet, press Ctrl and click (Command and click on Mac) on the measure "Number Of Reviews", to drag and drop it onto the dimension "Neighbourhood"

You should see a new object measure appearing, thanks to the Quick LOD feature.

Its expression is an LOD :

Testing Outcomes
Ok, but.. is that all? Yes: there is no additional help to the developer other than that, or at least that’s all I found.
A cool feature would have been a new dedicated LOD editor, to help the user defining the LOD scope and way to aggregate, but the new "Quick LOD" is not at all like that. It’s just an expression generator with no intelligence other than saving 5 seconds of development.
Also, in real life dashboards and data visualizations, often LOD expressions handle more than 1 dimension, so even the automatically generated expression will need to be edited and enhanced.
Also, it’s not clear to me why the FIXED LOD is the default Quick LOD type.
In conclusion, Data visualization developers will still need to learn LOD syntax, types, and logic to do their job properly, and there is almost no help from the new Quick LOD feature.
In a nutshell…
In this post I tested the new Tableau‘s feature "Quick LOD", that is supposed to help developers when dealing with Level Of Details expression. What I found is that Tableau tried to oversell it a bit, as the advantages it gives Vs manual LOD writing are quite negligible.
Thanks for reading.
Feel free to subscribe to my "Sharing Data Knowledge" Newsletter.

If you wish to subscribe to Medium, feel free to use my referral link https://medium.com/@maw-ferrari/membership : it costs the same for you, but it contributes indirectly to my stories.