Data Visualization
Data Visualization
Visualization is about matching a chart to a question and an audience. For comparison across categories, use a bar chart; for change over time, use a line chart; for distribution, use a histogram or box plot; for the relationship between two numeric variables, use a scatter plot; for part-to-whole with few slices, use a pie or stacked bar, but prefer bar charts when there are many categories. Heat maps show density across two dimensions; waterfall charts explain how a starting value moves to an ending value; treemaps show hierarchical proportions; gauges and bullet charts track a single KPI against a target.
Design matters as much as chart choice. Maximize the data-ink ratio, label axes, start bar charts at zero, sort categories meaningfully, and use color deliberately: sequential palettes for ordered magnitude, diverging palettes for values around a midpoint, and categorical palettes for unordered groups. Ensure contrast for color-blind viewers and add text or patterns when color alone carries meaning.
Traps appear in questions about truncated axes, which exaggerate differences; 3D charts, which distort; dual y-axes, which imply relationships that may not exist; pie charts with many slices; and dashboards crowded with detail. Also know the difference between a dashboard, which monitors ongoing operations, and a scorecard, which tracks KPIs against targets.
Worked example: a monthly revenue trend over two years belongs on a line chart with a zero-based y-axis for absolute values, but a line chart of percentage change may use a non-zero axis if clearly labeled. Rule of thumb: if the question emphasizes comparison of magnitude, use length (bars); if it emphasizes trend, use position over time (lines).