Determine data type before choosing a chart type

What it is

Task-first semantic typing states what relation a field represents and which comparisons the current question permits before selecting a bar chart, line chart, map, or scatterplot. It prevents a tool from treating every numeric column as a continuum, a date string as an unordered category, or a category code as a measure. “Before” names a decision dependency; during exploration, a provisional view may help revise the type judgment.

Why it happens

Chart types bundle marks, channels, and comparison assumptions. A connecting line suggests order and process, scatter position supports a relation between two quantities, and bar length implies magnitude accumulated from a reference. If semantic type is wrong, later color and label adjustments cannot remove the false relation. A database storage type specifies representation and available computation, not measurement scale, temporal meaning, missingness, or analytical purpose.

Studying it

Ask analysts to state the question, field meaning, unit, permissible operations, and forbidden interpretations before choosing a chart, then compare this workflow with default recommendations or immediate chart choice. Evaluate false relations in the output, task-answer accuracy, rework, and confidence. Include ambiguous codes, dates, ratings, percentages, and coordinates across several tasks.

Where it stops holding

Rapid exploration may generate a provisional view to expose data-quality or semantic problems, but the view is diagnostic evidence rather than proof that its inferred type is correct. Mixed fields, derived metrics, and uncertain measurements may require temporal, spatial, network, or statistical semantics beyond one three-way scale taxonomy. A type label also cannot replace domain judgment about comparability or causality.

Applying it

• Store field role, unit, scale type, task, and permissible transforms with chart configuration rather than only a column and chart name.
• Require semantic confirmation for numeric identifiers, time, percentages, and derived ratios. Keep unknowns pending instead of publishing a strong relational view automatically.
• Make recommendation systems explain which requested comparisons justify a channel, and preserve the source of user corrections.
• Before publication, check that the comparisons enabled by the graphic are exactly those required by the question and permitted by the measurement.

Related

• Same group: U1.08.1 Nominal, ordinal, and quantitative data need different encoding channels · U1.08.2 Encoding nominal data with a quantitative channel implies a nonexistent order · U1.08.3 Encoding quantitative data with a nominal channel discards magnitude differences · U1.08.5 The same field can take different types under different analytical tasks
• Adjacent: U2.01.1 Comparison, trend, distribution, composition, and relation each map to different charts · U2.01.2 Define the question before choosing the chart
• Search terms: semantic typing · chart selection · measurement scale