Learning Objectives

By the end of this lesson, learners should be able to:

• Identify major business visualization techniques.
• Select appropriate charts for different analytical questions.
• Explain the strengths and limitations of common visualization types.
• Interpret relationships, distributions and trends using graphical methods.
• Identify situations where a visualization may create misleading impressions.
• Evaluate charts based on analytical purpose and audience.

1. Introduction

Different analytical questions require different visualization techniques.

Selecting a chart should therefore begin with:

What decision or question is this visualization intended to support?

The analyst should then determine:

• What variables are involved?
• Are they numerical or categorical?
• Is time involved?
• Is the objective comparison, distribution, relationship or composition?
• Who will use the visualization?

2. Bar Charts

A bar chart compares values across categories.

Example:

Region	Revenue
Nairobi	40M
Mombasa	28M
Kisumu	18M
Nakuru	15M

A bar chart can make differences between categories easy to compare.

Best suited for:

• Category comparisons.
• Ranking.
• Discrete groups.

3. Column Charts

Column charts are essentially vertical bar charts.

They can be useful when:

• Categories are relatively few.
• Values need to be compared vertically.
• Time periods are being compared over a limited number of periods.

However, long category labels may be easier to read with horizontal bars.

4. Line Charts

Line charts are particularly useful for showing movement over time.

Example:

Monthly Revenue

January → February → March → April → May

They help reveal:

• Trends.
• Turning points.
• Seasonality.
• Short-term fluctuations.

Multiple lines can be used for comparison, but excessive series can make the chart difficult to interpret.

5. Scatter Plots

A scatter plot displays observations according to two numerical variables.

It can help identify:

• Positive relationships.
• Negative relationships.
• Weak relationships.
• Clusters.
• Outliers.

For example:

Training hours vs. employee productivity

A scatter plot may reveal whether the variables appear associated.

It does not by itself establish causality.

6. Histograms

A histogram displays the distribution of numerical observations across intervals.

It can help identify:

• Central concentration.
• Spread.
• Skewness.
• Multiple peaks.
• Potential unusual observations.

Histograms differ from bar charts because their bars represent intervals of numerical data rather than separate categories.

7. Box Plots

A box plot summarizes a distribution using key statistics such as:

• Median.
• Lower quartile.
• Upper quartile.
• Potential extreme observations.

Box plots are particularly useful when comparing distributions across several groups.

For example:

Customer spending across four regions.

They allow analysts to compare both central position and spread.

8. Pie Charts

Pie charts display parts of a whole.

They can be appropriate when:

• There are few categories.
• Categories represent a complete whole.
• Differences are relatively easy to distinguish.

However, pie charts become difficult to interpret when there are many categories or values are similar.

A bar chart may often provide more precise comparisons.

9. Stacked Bar Charts

Stacked bars show both:

• Total magnitude.
• Composition.

For example:

A company’s total revenue can be divided into:

• Product A.
• Product B.
• Product C.

Stacked charts can be useful, but comparing internal segments can become difficult when categories do not share a common baseline.

10. Area Charts

Area charts emphasize magnitude and trends over time.

They can communicate cumulative movement effectively.

However, overlapping areas can make comparisons difficult, particularly when several series are displayed.

11. Heat Maps

A heat map uses variations in visual intensity to represent values across two dimensions.

Examples include:

• Sales by region and month.
• Website activity by hour and day.
• Risk levels by business unit and risk category.

Heat maps are useful for identifying concentrations and patterns.

12. Geographic Maps

Maps can be useful when location is central to the analytical question.

Examples:

• Sales by county.
• Customer distribution.
• Branch performance.
• Delivery patterns.

Maps should not be used merely because geographic information is available. If location is not analytically important, another visualization may communicate the result more effectively.

13. KPI Cards

A KPI card highlights an important numerical indicator.

For example:

Revenue: KSh 125M

A stronger KPI card may include:

• Target.
• Previous period.
• Percentage change.
• Status.

KPI cards are useful for executive dashboards but provide limited detail on their own.

14. Tables

Tables remain valuable when precise numerical values are important.

Charts generally communicate patterns quickly, while tables support exact value retrieval.

A well-designed dashboard may therefore use both.

15. Choosing Between Charts

Consider the analytical objective:

Objective	Suitable Visualization
Compare categories	Bar chart
Show trend	Line chart
Examine distribution	Histogram
Compare distributions	Box plot
Examine relationships	Scatter plot
Show composition	Stacked bar
Show geographic patterns	Map
Highlight key metric	KPI card
Provide exact figures	Table

The choice should always reflect the analytical question.

16. Dual-Axis Charts

Dual-axis charts use two vertical scales.

They can be useful when variables have different units or magnitudes.

However, they can also create misleading visual relationships if scales are selected inappropriately.

For example:

Revenue and employee headcount

may appear strongly related simply because the two axes were visually aligned.

Dual-axis charts therefore require careful interpretation.

17. Small Multiples

Small multiples display the same type of chart repeatedly for different categories or groups.

For example:

• Revenue trend for Region A.
• Revenue trend for Region B.
• Revenue trend for Region C.

This can make comparisons easier than placing numerous lines on one chart.

18. Interactive Visualizations

Interactive tools allow users to:

• Filter.
• Drill down.
• Hover over values.
• Change time periods.
• Select categories.

Interactivity can increase analytical flexibility.

However, excessive interaction can also make dashboards confusing or hide important information.

19. Visualization Selection Framework

Before selecting a chart, ask:

Question 1

What am I trying to communicate?

Question 2

What type of variables do I have?

Question 3

What relationship should the audience see?

Question 4

What level of precision is required?

Question 5

Who will use the visualization?

This structured approach reduces arbitrary chart selection.

20. Executive Considerations

Executives generally require visualizations that allow rapid assessment of:

• Performance.
• Trends.
• Exceptions.
• Risks.
• Opportunities.
• Progress toward targets.

Technical complexity should not obscure the business message.

Lesson Summary

Different visualization techniques serve different analytical purposes.

Key choices include:

• Bar charts for category comparison.
• Line charts for trends.
• Scatter plots for relationships.
• Histograms for distributions.
• Box plots for comparing distributions.
• Stacked charts for composition.
• Heat maps for patterns across two dimensions.
• Maps for geographic analysis.
• KPI cards for critical metrics.
• Tables for precise values.

The strongest visualization is not necessarily the most sophisticated. It is the one that communicates the relevant evidence most accurately and efficiently.

References

1. Microsoft — Power BI
   Microsoft Power BI
2. Tableau — Data Visualization
   Tableau Data Visualization
3. NIST — Statistical Methods
   NIST Statistical Methods Handbook

Review Questions

1. What is the primary purpose of a bar chart?
2. When is a line chart appropriate?
3. What can a scatter plot reveal?
4. How does a histogram differ from a bar chart?
5. What information does a box plot provide?
6. When can a pie chart become ineffective?
7. What is the purpose of a stacked bar chart?
8. When should a map be used?
9. What is a KPI card?
10. Why can dual-axis charts be misleading?
11. What are small multiples?
12. Why should visualization selection begin with the analytical question?