Data Visualization

Turning Numbers Into Understanding

Visualizations reveal patterns that numbers hide. They communicate to others what took you hours to discover. Done well, they make data intuitive.

Why Visualize

Pattern Recognition

Humans process visual information faster than numbers. Charts reveal trends, outliers, and relationships instantly.

Communication

A good chart conveys findings faster and more memorably than tables.

Exploration

Visualizing data helps you understand it better during analysis.

Choosing the Right Chart

For Distribution (One Variable)

Histogram: Shows frequency distribution of continuous data.

  • How are values spread?
  • Where do most values fall?
  • Any outliers or gaps?

Bar Chart: Shows counts or values for categories.

  • Compare quantities across categories
  • Rank or order matters

Box Plot: Shows distribution summary compactly.

  • Compare distributions across groups
  • Show median, quartiles, outliers

For Relationships (Two Variables)

Scatter Plot: Two continuous variables.

  • Correlation
  • Clusters
  • Outliers

Line Chart: One variable over time.

  • Trends
  • Seasonality
  • Change over time

Grouped Bar Chart: Categories compared across another dimension.

  • Product sales by region
  • Metrics by time period

For Composition

Pie Chart: Parts of a whole (use sparingly).

  • Only for a few categories
  • Only when parts sum to meaningful whole

Stacked Bar Chart: Composition across categories.

  • Category breakdown over time
  • Component comparison

100% Stacked Bar: Proportional composition.

  • When absolute size varies but proportions matter

For Comparison

Bar Chart: Compare values across categories.

  • Horizontal bars for many categories
  • Sort by value for easier reading

Bullet Chart: Actual vs. target.

  • Performance against goals

Small Multiples: Same chart repeated.

  • Compare patterns across segments

Principles of Good Visualization

Show the Data

The data should be the star. Don't obscure it with decoration.

Minimize Chartjunk

Remove unnecessary elements:

  • 3D effects (distort perception)
  • Excessive gridlines
  • Unnecessary decoration
  • Redundant labels

Use Appropriate Scales

  • Start bar charts at zero
  • Don't manipulate axes to exaggerate
  • Use consistent scales when comparing

Label Clearly

  • Clear title stating the insight
  • Axis labels with units
  • Legend only when needed
  • Data labels where helpful (not everywhere)

Consider Color

  • Use color purposefully
  • Consistent color meaning
  • Accessible for color blindness
  • Don't use color alone to convey information

Tell a Story

Good visualizations have a point. What should the viewer take away?

Common Visualization Mistakes

Pie Charts with Too Many Slices

Pie charts work poorly with more than 5-6 categories. Use bar charts instead.

Truncated Axes

Starting a bar chart at a value other than zero exaggerates differences.

3D Effects

3D adds nothing and distorts perception. Avoid it.

Dual Axes

Two y-axes can mislead by manipulating scale. Use carefully or not at all.

Rainbow Colors

Too many colors confuse. Use a limited, meaningful palette.

Overloaded Charts

Too much data in one chart. Split into multiple charts.

Creating Visualizations

In Spreadsheets

Excel and Google Sheets can create basic charts effectively:

  1. Select data
  2. Insert chart
  3. Choose type
  4. Format and label
  5. Refine

With AI

Describe what you want:

  • "Create a bar chart showing sales by region"
  • "Show me the trend of monthly revenue over time"
  • "Compare distributions of order value by customer segment"

AI can generate code or describe how to create visualizations.

AI Prompt: Visualization Recommendation

Help me choose the right visualization.

My data: [Describe what you have]
What I want to show: [The insight or comparison]
Audience: [Who will see this]

Recommend:
1. Best chart type and why
2. How to structure the data for this chart
3. Key design decisions
4. What labels and annotations to include

AI Prompt: Chart Creation

Help me create a visualization.

Data to visualize: [Paste or describe data]
Chart type: [If you know it, or "recommend"]
Key message: [What should viewer take away]
Tool: [Excel, Google Sheets, etc.]

Please provide:
1. Step-by-step creation instructions
2. Recommended formatting
3. Suggested labels and title
4. How to improve clarity

What's Next

Beyond description, let's find relationships.

Next chapter: Finding relationships — correlation and comparison.

Descriptive StatisticsFinding Relationships