5.2 View Granularity, Aggregations & Measure Names/Values

5.2 View Granularity, Aggregations & Measure Names/Values

In Tableau Desktop, every visualization is governed by two fundamental concepts: View Granularity (also known as the Viz Level of Detail or Viz LOD) and Aggregation. Understanding how dimensions slice data into discrete marks, how measures aggregate numerical values within those marks, and how Tableau's special generated fields—Measure Names and Measure Values—behave on axes and shelves is core knowledge for the Tableau Desktop Specialist exam.

---

Understanding Viz Level of Detail (Viz LOD)

The Viz Level of Detail (Viz LOD) defines the specific level of data abstraction displayed in a visualization. It determines the number of individual visual marks (bars, points, lines, pie slices) rendered on the canvas, as well as the exact granularity at which measures are aggregated.

Dimensions Determine Marks; Measures Do Not

A common misconception on the exam is that adding measures increases the number of marks in a view. In Tableau, dimensions establish granularity, while measures evaluate quantitative values within that granularity.

1. Adding Dimensions Increases Granularity: When you place a dimension (such as Region or Category) on the Rows, Columns, Color, Size, Label, Detail, or Shape shelves on the Marks card, Tableau breaks down the underlying data into finer slices. Each distinct combination of dimension values forms a separate mark.
2. Adding Measures Does Not Change Granularity: Dragging a measure (such as Sales or Profit) onto Rows or Columns does not alter the total mark count. Instead, it creates or updates a quantitative axis that displays the aggregated total for each existing dimension slice.
3. The Detail Shelf: Placing a dimension on the Detail shelf alters the Viz LOD by creating additional marks without adding visual encoding like color, size, or shape. This is particularly useful for scatter plots or reference line calculations where detailed granularity is required without cluttering visual channels.

If you remove all dimensions from a worksheet and drag Sales to Rows, Tableau displays a single mark representing the grand total SUM(Sales) across the entire dataset. Adding Category (with 3 members) splits that single mark into 3 marks. Adding Sub-Category (with 17 members) further splits the view into 17 marks.

---

Aggregation Mechanics & Mathematical Operators

When continuous or discrete measures are brought into a view containing dimensions, Tableau automatically aggregates raw data rows. Aggregation combines multiple row values into a single summary number for each mark.

Key Aggregation Functions

• SUM([Field]): Calculates the arithmetic total of all non-null numeric values within the slice.
• AVG([Field]): Computes the arithmetic mean (SUM divided by COUNT).
• COUNT([Field]): Counts the total number of non-null data rows in the partition.
• COUNTD([Field]): Calculates the count of distinct (unique) values. COUNTD requires more computational memory than COUNT because Tableau must evaluate uniqueness across rows.
• MIN([Field]) / MAX([Field]): Returns the minimum or maximum numeric, date, or string value in the group.
• MEDIAN([Field]): Identifies the middle value in a sorted distribution.
• ATTR([Field]): The Attribute function. Returns the literal value of a field if every row in the partition contains the exact same single value; otherwise, it returns an asterisk (*). Mathematically, ATTR(X) is evaluated as IF MIN(X) = MAX(X) THEN MIN(X) ELSE '*' END.

Row-Level Calculations vs. Aggregated Calculations

Understanding the mathematical difference between row-level calculations and aggregated calculations is vital for exam success:

// Row-Level Calculation (Calculated BEFORE aggregation)
[Profit Ratio Row Level] = [Profit] / [Sales]

// Aggregated Calculation (Calculated AFTER aggregation)
[Profit Ratio Aggregated] = SUM([Profit]) / SUM([Sales])

• Row-Level Calculation ([Profit] / [Sales]): Tableau divides profit by sales for each individual transaction row first. When brought into the viz, Tableau sums those pre-calculated ratios (SUM([Profit Ratio Row Level])). This yields an incorrect, mathematically invalid profit margin percentage.
• Aggregated Calculation (SUM([Profit]) / SUM([Sales])): Tableau first sums all profit across all rows in the mark, then sums all sales across all rows in the mark, and finally divides total profit by total sales. This produces the mathematically correct overall profit margin.

Disaggregating Data

By default, Tableau aggregates measures (Analysis > Aggregate Measures is checked). If you uncheck Aggregate Measures, Tableau disaggregates the data, plotting every raw transaction row as an individual mark. Disaggregation is primarily used when analyzing raw data distribution in scatter plots.

---

Measure Names and Measure Values Generated Fields

When working with multiple measures in a single view, Tableau automatically introduces two special generated fields shown in italicized text in the Data pane: Measure Names and Measure Values.

What are Measure Names and Measure Values?

• Measure Names: A discrete dimension that contains the literal string names of all measures in your data source (e.g., "Sales", "Profit", "Discount", "Quantity").
• Measure Values: A continuous measure that contains the numerical data values of all measures listed in the Measure Values card.

How Tableau Creates a Blended Axis

When you drag a second measure directly onto the axis of an existing measure in a view, Tableau automatically builds a Blended Axis visualization:

1. Tableau places Measure Names on the Columns (or Rows) shelf and on the Color card.
2. Tableau replaces the individual measure field on Rows (or Columns) with the Measure Values single continuous axis field.
3. Tableau opens the Measure Values card shelf below the Marks card, listing all measures sharing that unified axis.

Inside the Measure Values card, each measure can have its own independent aggregation method (e.g., SUM(Sales), AVG(Profit)), but all measures share the same quantitative axis scale.

---

Comparative Analysis Table: Blended Axis vs. Dual Axis

---

Key Rules for the Exam

1. Placing a dimension on any Marks card shelf (Color, Size, Detail, Shape, Label) increases the view granularity.
2. ATTR() returns an asterisk (*) when a dimension contains multiple distinct values within a single aggregated mark slice.
3. Blended Axis views combine multiple measures onto a single axis using Measure Names and Measure Values, supporting unlimited measures.
4. Dual Axis views plot two measures on two separate axes, allowing independent mark types (e.g., combining a Bar chart with a Line chart).