Benefits and limitations of data visualization

Learning outcome statements

The learning outcome statements relevant for this section are:

1. Utilize table and graph design best practices to avoid distortion in the communication of complex information
2. Evaluate data visualization options and select the best presentation approach (e.g., histograms, box plots, scatterplots, dot plots, tables, dashboards, bar charts, pie charts, line charts, bubble charts) for a given scenario
3. Understand the benefits and limitations of visualization techniques
4. Communicate results, conclusions, and recommendations in an impactful manner using effective visualization techniques

This section covers LOS 3 - the benefits and limitations of visualization techniques. Chart selection, design best practices, and communicating results are covered in the statistical and business data visualization methods sections that follow.

The benefits and limitations of visualization techniques

Benefits of data visualization

• Enhanced understanding:
  Data visualization simplifies complex datasets, making it easier to spot trends, patterns, correlations, outliers, and anomalies than scanning raw numbers alone.
• Improved decision-making:
  Clear visual representations of data help businesses and analysts make informed decisions by quickly identifying key insights.
• Better communication:
  Visualized data is more engaging and accessible, improving communication between analysts, stakeholders, and decision-makers.
• Increased retention and engagement:
  Well-designed visuals are often easier to grasp and remember than dense tables or text, which makes them useful in presentations and reports.

Limitations of data visualization

• Potential for misrepresentation:
  Poor design choices can distort how data is interpreted - common techniques include a truncated or non-zero axis baseline, inconsistent or dual axes, 3-D or area effects that exaggerate size, cherry-picked time windows, and omitted context. A visual pattern, such as a trend in a scatterplot, shows correlation, not causation - the pattern may be driven by an omitted variable rather than a direct cause-and-effect relationship.
• Over-simplification of data:
  While visualization makes data more accessible, it can sometimes strip away critical details or nuances needed for deeper analysis.
• Bias in presentation:
  The way data is visualized can introduce bias, either intentionally or unintentionally, influencing the audience’s perception.
• Complexity in large datasets:
  While visualization aids analysis, very large and complex datasets may require multiple visualizations or advanced tools to fully convey insights.
• Dependence on proper interpretation:
  Data visualizations are only as effective as the user’s ability to interpret them correctly; misinterpretations can lead to incorrect conclusions.

Example: Truncated axis

Quarterly revenue rises from $98 million to $102 million, about a 4% increase. A bar chart with the y-axis starting at $0 shows two bars of nearly equal height. The same data plotted with the y-axis starting at $95 million makes the increase look several times larger than it really is.

Takeaway: Always check where a bar chart’s y-axis starts before drawing conclusions - a non-zero baseline can make a small change look dramatic.