Using LOD Expressions to Ask Deeper Questions
Tableau is a great tool for data visualization. The challenge in data analysis often involves taking a question that is easy to articulate out loud, then answering that question with data. The analysis frequently requires comparing or referencing data at multiple different aggregation levels.
The primary focus of the visualization may be at one level, but the question may reference another level. You might, for example, be visualizing sales revenue at a country level on a map, but you want to compare those sales to the global sales number. And that’s where Level of Detail (LOD) expressions come in.
In Tableau, the main visualization aggregation level is referred to as its “level of detail.” LOD expressions allow us to look beyond the visualization level of detail.
The data in the visualization is often a result of filtering the data from the data source. LOD expressions can look at the data before it is filtered, which enables more powerful analyses.
First, let’s consider some typical analytical questions. They are easy to express in everyday language:
What is the largest deal closed per salesperson, and the average of these deals per country?
What is the cumulative number of customers acquired per day by market?
What is the difference in year-to-date profit of this year versus last year, relative to the last day the data was refreshed?
There is a clear focus in each of these questions, but they also reference another context.
There are several ways of referencing different levels of detail. The first example looks down a level of detail, the second example looks outside the current level of detail, and the third example looks up a level of detail.