With little data and few dimensions, tables are denser

What it is

Three options against four dimensions: a 3×4 table delivers every true value in a palm-sized patch of layout; the same data as a chart spends most of its space on axes, legend, and margins — and every number becomes an estimate meanwhile. "Small data goes in a table" is not laziness but the result of a density calculation: the data's scale cannot amortize the chart's structural overhead (axes, legend, grid, titles), while every cell of a table carries true data. Only when data grows large and dimensions multiply does the balance flip.

Why it happens

Charts carry fixed overhead: whatever the data volume, axes and ticks, legend, titles, and sensible margins must all appear — a line chart of 5 points and one of 50 cost nearly the same, so density amortizes only as data grows. A table's every row and column is the data itself, its only overhead being the header, so at small scale its density is unbeatable; but it grows serially (one more row, one more scan), and past a hundred rows the attention cost of lookup and comparison climbs linearly while the chart's parallel perception takes over. The two density curves cross near "the rows that fit one screen"; before the crossing, the table is the heavier representation.

Where it stops holding

"Small" is judged not by absolute row count but by the scale used in one act of reading: within one screen (roughly 20–30 rows) is table country; beyond it, charts or interactive filtering take over. "Few dimensions" matters equally: four or five is the table's comfort zone; more and the table sprawls horizontally, with wide-table trade-offs entering the multidimensional-representation family's territory (parallel coordinates and kin). In report pages a table also carries citability (numbers one can point to) — a side benefit beyond density.

Applying it

• With a dozen-ish rows, four or five dimensions, and any value-retrieval need, default to a table; do not chart "to look more analytical."
• Give the table layout alignment resources: right-aligned numerals, tabular figures, one unit and quantity per column.
• Verification: count the candidate chart's data-ink share (data-element area / plot area); under half with rows fitting one screen means switch to a table.

Related

• Same group: U2.16.1 Tables beat charts when exact values are needed · U2.16.2 Charts beat tables for discovering patterns and comparing trends · U2.16.4 Sparklines in tables serve readout and trend at once · U2.16.5 Row ordering does the same work as category ordering in charts
• Nearby: U2.14.1 Parallel coordinates render each record as a polyline across axes · U1.14.1 Non-data elements consume attention while carrying no information
• Search terms: table density · small data · data-ink · tabular layout