Understanding Data Types

Before you choose a chart, you need to know what kind of data you have. Statisticians sort data into four "levels of measurement" — nominal, ordinal, interval, and ratio — and the level tells you what comparisons are meaningful and, crucially, which chart will read clearly. Get the data type right and the chart almost picks itself; get it wrong and you end up averaging things that cannot be averaged or plotting a line through categories that have no order.

The four levels form a ladder. Each step up allows everything the level below allowed, plus one new kind of comparison. The first two levels are categorical — they label things into groups. The last two are continuous (numeric) — they measure how much.

Nominal — unordered categories

Nominal data sorts items into named groups with no inherent order. Colors are the classic example: red, blue, and green are different, but none is "more" than another. Other examples are country names, product categories (Category A, Category B, Category C), or a yes/no answer. You can count how many items fall in each group, and you can check whether two values are the same or different — but you cannot rank them or do arithmetic on the labels themselves. There is no "average color." The only summary that makes sense is a count per category, or the mode (the most common category).

Ordinal — ordered categories

Ordinal data is categorical too, but the categories have a meaningful order. Sizes such as small, medium, and large are ordinal: large is bigger than medium, which is bigger than small. So are satisfaction ratings (poor, fair, good, excellent) or finishing positions (1st, 2nd, 3rd). You can rank ordinal values and find a median, but the gaps between ranks are not guaranteed to be equal. The jump from "good" to "excellent" may not be the same size as the jump from "poor" to "fair," so averaging the ranks can mislead. Treat ordinal data as ordered labels, not as numbers.

Interval — numbers without a true zero

Interval data is genuinely numeric, with equal, meaningful gaps between values — but it has no true zero. Temperature in degrees Celsius is the standard example: the difference between 10° and 20° is the same as between 20° and 30°, so you can add and subtract and find a meaningful mean. But 0°C does not mean "no temperature," and that missing true zero means ratios break down — 20°C is not "twice as warm" as 10°C. Calendar years behave the same way: the year 2000 is not twice the year 1000. Interval data supports differences and averages, but not ratios.

Ratio — numbers with a true zero

Ratio data is numeric with equal gaps and a true zero that means "none of the quantity." Length, weight, age, duration, and any count (number of items, visits, sales) are ratio data. Because zero is real, every kind of arithmetic works, including ratios: 20 kg really is twice 10 kg, and 0 items really means nothing is there. This is the richest level — you can count, rank, average, and compare multiples. Most measured and counted quantities you will chart are ratio data, which is why variance, standard deviation, and other numeric summaries apply naturally to them.

Which chart suits which data type

This is where the four levels pay off. The data type points straight at a chart:

• Nominal & ordinal (categorical) → a bar chart of the count or total per category. For a few parts of one whole, a pie chart can work. For ordinal data, keep the bars in the natural order (small → large), not sorted by height.
• Ratio counts of one variable → a histogram, which groups numeric values into ranges and shows their distribution.
• Interval or ratio over time → a line chart, because the x-axis is a continuous, evenly spaced range.
• Two numeric variables together → a scatter plot to reveal whether they are related.

The single biggest mistake is treating one type as another: drawing a line chart across nominal categories (implying a trend that does not exist), or making a bar chart of every individual numeric reading instead of summarizing it into a histogram. When you are unsure, our how to choose a chart guide walks the decision from your data type to the right picture, and the full chart-type reference shows each option with examples.

Once you can name your data type, charting stops being guesswork. Identify whether each column is nominal, ordinal, interval, or ratio first — and the rest of the work, from summarizing to choosing the chart, follows from that.