7.1 Introduction

Data visualization transforms complex datasets into clear, actionable narratives (Few 2009). Charts and graphs highlight trends, anomalies, and patterns that would be difficult to spot in raw tables, enabling faster and more confident decisions. Visualization also makes data approachable for non-technical stakeholders, promoting a data-driven culture across the organization.

Visualization serves two distinct purposes in BI. Exploratory visualization is for the analyst — quick, iterative plots created during data analysis to discover patterns, test hypotheses, and identify anomalies. These visualizations are often rough and informal; their purpose is insight, not presentation. Explanatory visualization is for the audience — carefully designed charts that communicate a specific finding or recommendation to stakeholders. These emphasize clarity, simplicity, and narrative impact (Tufte 2001). Knowing which mode you are in matters: an exploratory scatter plot that reveals an interesting trend may need to be completely redesigned as an explanatory chart before it is ready for a presentation.

Choosing the right visualization starts with understanding the type of relationship you want to show (Heer and Shneiderman 2012):

  • Correlation: The association between two variables — do they move together, in opposite directions, or independently? Visualized with scatter plots.
  • Distribution: How data points spread across a range of values — what is the center, spread, and shape? Visualized with histograms and box plots (Tufte 2001).
  • Comparison: Differences and similarities across categories or over time. Visualized with bar charts and line graphs (Yau 2011).
  • Composition: How parts contribute to a whole — what proportion does each component represent? Visualized with stacked bar charts and pie charts (Cairo 2012).
  • Complex relationships: Multi-variable interactions that require heat maps, network diagrams, or multi-dimensional plots to unravel (Few 2009).

This chapter is organized around three layers. First, the why — storytelling with data and matching chart types to questions. Second, the principles — design guidelines that separate effective visualizations from misleading ones. Third, the how — the Grammar of Graphics framework, its implementation in ggplot2, and how it compares to other visualization tools (Wickham 2016). The chapter concludes with a look at how AI is beginning to assist with visualization tasks.