4 Graphical Representation

What This Chapter Covers

Graphical representation is the systematic use of charts, diagrams, and visual forms to encode data for human interpretation. In this chapter, you will learn the full taxonomy of chart types used in data visualization, the encoding principles that make each type effective, and how to build the most important chart types, bar charts, line charts, area charts, pie charts, treemaps, and maps, in Tableau. You will also learn the design rules that govern when each type is appropriate and when it is misleading.

flowchart TD
    A[Graphical Representations] --> B[Comparison]
    A --> C[Trend over Time]
    A --> D[Part-to-Whole]
    A --> E[Distribution]
    A --> F[Relationship]
    A --> G[Geographic]
    B --> B1[Bar chart <br> Lollipop chart]
    C --> C1[Line chart <br> Area chart]
    D --> D1[Pie chart <br> Treemap <br> Stacked bar]
    E --> E1[Histogram <br> Box plot <br> Violin plot]
    F --> F1[Scatter plot <br> Bubble chart]
    G --> G1[Symbol map <br> Filled map <br> Flow map]
    style A fill:#e3f2fd,stroke:#1976D2
    style B fill:#f3e5f5,stroke:#7B1FA2
    style C fill:#fff9c4,stroke:#F9A825
    style D fill:#fce4ec,stroke:#C62828
    style E fill:#e8f5e9,stroke:#388E3C
    style F fill:#e0f7fa,stroke:#0097A7
    style G fill:#fff3e0,stroke:#E65100

4.1 A Taxonomy of Chart Types

Organising Charts by Analytical Purpose

Every chart type is designed to answer a specific class of question. Before selecting a chart, identify which of the following analytical tasks you need to perform:

Analytical Task	Question Answered	Primary Chart Types
Comparison	How do these values differ across categories?	Bar chart, lollipop chart, dot plot
Trend	How has this value changed over time?	Line chart, area chart, slope chart
Part-to-whole	What share does each part contribute to the total?	Pie chart, treemap, stacked bar chart, waterfall chart
Distribution	How are these values spread across a range?	Histogram, box plot, violin plot, strip plot
Relationship	Is there a pattern between two variables?	Scatter plot, bubble chart, heatmap
Geographic	Where are these values located?	Symbol map, filled map, flow map
Flow	How does a value move between states?	Sankey diagram, waterfall chart

This taxonomy is the foundation of chart selection. Master it, and you will never default to the wrong chart type again.

4.2 Bar Charts: The Workhorse of Comparison

Why Bar Charts Are the Most Reliable Comparison Chart

The bar chart encodes quantitative values as the length of rectangular bars. Because length is one of the most accurately perceived pre-attentive attributes (as established in Chapter 1), bar charts are uniquely reliable for categorical comparison. Viewers can accurately judge the difference between two bar lengths even when the bars are not adjacent.

Key design rules: 1. Always start the Y-axis (or X-axis for horizontal bars) at zero. Truncating the axis makes small differences appear large, a common form of chart manipulation. 2. Sort bars by value (descending) unless the category has a natural order (e.g., months, age groups). 3. Use horizontal bar charts when category labels are long, they are easier to read than angled or rotated labels on a vertical chart. 4. Avoid 3D bars, they distort the length encoding by adding a misleading depth dimension.

How To: Building a Sorted Bar Chart in Tableau

Drag Sub-Category to the Rows shelf.
Drag Sales to the Columns shelf. Tableau creates a horizontal bar chart.
To sort by value: click the sort descending button on the toolbar (the icon with bars and a downward arrow), or right-click the Sub-Category pill and select Sort > Field > Sales > Descending.
Drag Category to the Colour shelf to colour bars by category.
Right-click the X-axis and select Edit Axis to rename it “Total Sales (USD).”
To add data labels, click the Label shelf on the Marks card, tick Show mark labels, and set Alignment to right.

[Insert screenshot of the completed sorted horizontal bar chart with Category colour coding and data labels]

Grouped and Stacked Bar Charts

When you need to compare values across two categorical dimensions simultaneously, you have two options:

Grouped bar chart: Place one dimension on Rows/Columns and a second dimension on the Colour shelf. Each category has a cluster of bars, one per sub-category. Best for comparing within groups.

Stacked bar chart: Stack multiple series inside each bar. Best for showing total magnitude alongside composition. Change the mark type to Bar and place both dimensions using the Colour shelf.

100% stacked bar chart: Normalises all bars to 100%, making it easy to compare proportions across categories while hiding absolute values. Right-click the measure axis and select Add Reference Line > Constant > 100%, or use a Table Calculation (% of Total).

When to Use Lollipop Charts Instead of Bar Charts

A lollipop chart replaces the filled bar with a thin line and a circle at the end. It conveys the same information as a bar chart but with less ink, which can improve readability when you have many categories. To create one in Tableau: change the mark type to Line, add a second layer using a dual-axis with mark type Circle, and synchronise the axes.

4.3 Line Charts: The Standard for Trends Over Time

When to Use a Line Chart

A line chart connects data points with a line to show how a measure changes over a continuous dimension, almost always time. The line implies continuity between points, which is why line charts should only be used when the X-axis is truly continuous (time, distance, temperature) and not for categorical comparisons.

Key design rules: 1. Use line charts for time-series data only. For categorical X-axes, use bar charts. 2. Limit the number of lines to five or fewer. Beyond five, colours become hard to distinguish and the chart becomes a “spaghetti chart.” 3. Label lines directly at the right end rather than relying on a colour legend, this reduces eye movement and makes the chart faster to read. 4. Use a secondary axis (dual axis) with caution and always synchronise the axes to prevent misleading comparisons.

How To: Building a Multi-Line Chart in Tableau

Drag Order Date to the Columns shelf. Right-click the date pill and set the granularity to Month (continuous).
Drag Sales to the Rows shelf.
Drag Category to the Colour shelf, Tableau creates one line per category, each a distinct colour.
To label lines at the endpoint: click the Label shelf, tick Show mark labels, set Marks to Label to Most Recent.
To add forecast lines: open the Analytics pane and drag Forecast onto the view.

[Insert screenshot of a three-line chart (one per Category) with month-level granularity, forecast lines, and endpoint labels]

Area Charts: Adding Volume to the Line

An area chart fills the space between a line chart and the zero baseline with a solid colour or a gradient. The fill emphasises cumulative volume rather than individual point values, making area charts best suited for showing the total magnitude of a metric over time.

Stacked area charts stack multiple series, showing both individual contributions and the total. Change the mark type from Line to Area on the Marks card. Use opacity (80% or less) so that underlying series remain visible.

4.4 Pie Charts and Treemaps: Part-to-Whole Composition

Pie Charts: Use Sparingly and Correctly

A pie chart encodes part-to-whole relationships as angles and arc lengths within a circle. Despite its ubiquity in business reporting, the pie chart is one of the least perceptually accurate chart types because humans judge angles and areas less accurately than lengths.

When pie charts are acceptable: - You have two to four categories, and the purpose is to emphasise the dominant category’s share. - The audience is a general (non-technical) audience where familiarity with pie charts outweighs its perceptual limitations.

When to avoid pie charts: - More than five categories, use a bar chart instead. - When precise comparisons between similarly-sized slices are needed. - Never use 3D pie charts, the perspective distortion makes proportions impossible to judge accurately.

How To: Creating a Pie Chart in Tableau

Open a new worksheet and change the mark type to Pie in the Marks card dropdown.
Drag Sales to the Angle shelf (this controls slice size).
Drag Category to the Colour shelf.
Drag Category and Sales to the Label shelf and format to show both the category name and the percentage.
Resize the chart to be small, pie charts are best displayed at small sizes where the dominant slice is immediately apparent.

[Insert screenshot of a Tableau pie chart with three categories, percentage labels, and a clean colour palette]

Treemaps: Hierarchical Part-to-Whole

A treemap displays hierarchical data as nested rectangles, where each rectangle’s area is proportional to a quantitative measure. Treemaps are superior to pie charts when you have many categories (10 or more) or a hierarchical structure (e.g., Category > Sub-Category).

How to read a treemap: Larger rectangles represent larger values. Colour can encode a second measure (e.g., profit ratio) using a diverging palette. The hierarchical nesting shows the parent-child relationship between category levels.

How to create a treemap in Tableau: 1. Hold Ctrl and select Category, Sub-Category, and Sales. 2. In Show Me, click Treemap. Tableau nests sub-categories within categories. 3. Drag Profit to the Colour shelf to create a dual-encoding treemap (size = sales, colour = profit).

When to Choose Treemap Over Stacked Bar

Use a treemap when you have more than 10 sub-categories and want to show hierarchical structure and relative size simultaneously. Use a stacked bar chart when you want to compare totals across groups while showing composition, the stacked bar makes cross-group comparison easier because bars are aligned on a common baseline.

4.5 Geographic Maps in Tableau

When Geography Adds Value

A geographic map should be used only when location is a meaningful variable, not simply because your data contains location fields. A map is the right choice when the spatial distribution of the data is the primary insight: where are our customers concentrated? Which regions have the highest defect rates? Where are our delivery delays occurring?

Tableau supports three primary map types: - Symbol map, Plots circles (or other shapes) at geographic coordinates. The size and colour of the symbol encodes quantitative values. Best for point data (cities, stores, events). - Filled map (choropleth), Fills geographic regions (countries, states, postcodes) with colour to encode a quantitative variable. Best for showing regional rates or aggregated values. - Flow map, Draws lines between origin and destination points, with line width encoding flow volume. Built in Tableau using the Path shelf on line marks.

How To: Creating a Symbol Map in Tableau

Drag State (or Country) to the Details shelf. Tableau automatically recognises geographic fields and switches to a map background.
Drag Sales to the Size shelf, circles scale by total sales per state.
Drag Profit to the Colour shelf with a diverging palette (negative profit = red, positive profit = blue).
To switch from a symbol map to a filled map: change the mark type from Circle to Filled Map in the Marks card dropdown.
Adjust the Map Layers (Map menu > Map Layers) to add or remove street names, coastlines, and country borders.

[Insert screenshot of a symbol map of the United States showing Sales as circle size and Profit as colour, with the Marks card visible]

The Filled Map’s Hidden Distortion: Population and Area

Filled maps (choropleths) are distorted by the geographic size of regions. A large, sparsely populated state (e.g., Alaska) visually dominates the map even if it contributes minimally to the business. Always consider whether geographic area correlates with your data, and if it does not, add a supplementary bar chart showing the same values ranked by size to provide accurate context.

4.6 Choosing the Right Chart: A Decision Framework

The Chart Selection Decision Tree

Use this framework every time you are about to build a new chart:

Step 1, Identify the analytical task: Comparison, trend, distribution, relationship, composition, or geography?

Step 2, Count the variables: How many dimensions and measures are involved?

Step 3, Check the data types: Are the variables continuous, ordinal, or nominal?

Step 4, Consider the audience: Is this for EDA (allow complexity) or communication (maximise simplicity)?

Task	Variables	Recommended Chart
Comparison	1 dimension + 1 measure	Horizontal bar chart
Comparison	2 dimensions + 1 measure	Grouped or stacked bar chart
Trend	Date + 1 measure	Line chart
Trend	Date + multiple measures	Multi-line chart
Distribution	1 continuous measure	Histogram
Distribution	1 continuous + 1 categorical	Box plot
Relationship	2 continuous measures	Scatter plot
Composition	1 dimension + 1 measure (few categories)	Pie chart
Composition	Hierarchy + 1 measure (many categories)	Treemap
Geographic	Location + 1 measure	Symbol or filled map

4.7 Summary

Key Concepts at a Glance

Chart Type	Encodes	Best For
Bar chart	Length	Categorical comparison
Line chart	Position over time	Trends over continuous X
Area chart	Area under line	Cumulative volume over time
Pie chart	Angle/arc	Part-to-whole (up to 5 categories)
Treemap	Area + colour	Hierarchical composition
Symbol map	Position + size + colour	Geographic point data
Filled map	Fill colour	Geographic rates or aggregates

Applying This in Practice

In Tableau, the Show Me panel is a useful starting point, but it should not be the final word on chart selection. Show Me recommends chart types based on the number and type of fields you have selected, but it does not know the analytical question you are trying to answer. Use Show Me to explore options quickly, then apply the decision framework above to confirm whether the suggested chart is the right one for your specific purpose.