Choosing and setting out a graph
Every graph must be easy for a stranger to read. Whatever type you pick, include a clear title, a scale with equal steps, labelled axes with units (for example "Rainfall (mm)"), and a key if there is more than one set of data. Use annotations, short notes on the graph, to point out something important such as a peak.
Key terms:
- Scale: the equal steps marked along an axis, such as 0, 10, 20, 30.
- Annotation: a short label added to a graph to point out or explain a feature.
- Frequency: how many times a value, or a value in a class, occurs.
- Best-fit line: a straight line drawn through the middle of the points on a scatter graph.
- Dispersion: how spread out a set of values is.
| If the data is... | Use a... |
| Separate categories, such as countries | Bar graph |
| Change over time | Line graph |
| Parts of a whole | Pie chart or doughnut graph |
| One measurement split into equal classes | Histogram |
| Two sets of data that may be linked | Scatter graph |
| The spread of a set of values | Dispersion graph |
| A quick, simple picture of amounts | Pictogram |
Bar graphs and line graphs
📊 Bar graphs
Use bars for separate categories, such as countries or farm types. Bars have gaps between them, are all the same width, and start from zero. Vertical bars are the usual choice. Horizontal bars suit long labels. A divided bar splits each bar into sections that add up to the whole, which shows how a total is made up.
📈 Line graphs
Use a line for data that changes continuously, usually over time. Put time on the horizontal axis, plot each point, then join the points. A line graph shows a trend very clearly and can show more than one line, each with its own key.
Pie charts and doughnut graphs
A pie chart shows parts of a whole as slices of a circle. A circle has 360 degrees, so each 1% is 3.6 degrees.
Worked example
Land use in a made-up town: housing 45%, industry 25%, open space 20%, shops 10%. Angle = percentage x 3.6. Housing: 45 x 3.6 = 162°. Industry: 25 x 3.6 = 90°. Open space: 20 x 3.6 = 72°. Shops: 10 x 3.6 = 36°. Check: 162 + 90 + 72 + 36 = 360°. Working backwards, a slice of 54° is 54 ÷ 3.6 = 15%.
Draw the slices with a protractor, start at the top and work clockwise, and label each slice or give a key. A doughnut graph is a pie chart with a hole in the middle. It is drawn in the same way. Several rings can be nested, one for each year, so changes can be compared, and the hole can hold a total. Pie charts and doughnut graphs are poor when there are many tiny slices.
Pictograms
A pictogram uses a small picture to stand for a number. The key must say what one symbol means, for example one bicycle = 20 bicycles. Half a symbol is half the number, so 50 bicycles is two and a half symbols. Pictograms are quick to read but not good for exact values.
Histograms
A histogram shows how continuous data is spread across classes. The classes must be equal in width, such as 0 to under 5 minutes, 5 to under 10 minutes, and so on. The bars touch because the data runs on without a break, and the vertical axis shows frequency. The tallest bar is the most common class.
A bar graph has gaps and shows separate categories. A histogram has no gaps and shows one measurement split into classes.
Scatter graphs
A scatter graph tests whether two sets of data are linked. Each point is one pair of values, for example distance from the town centre and house price. Draw a best-fit line by eye through the middle of the points, with about the same number of points above and below it. It does not have to touch every point.
As in Comparing Countries with Development Indicators, the best-fit line shows the type of correlation. An upward slope is positive, a downward slope is negative (house price usually falls as distance from the centre rises), and points with no pattern mean no correlation.
The nearer the points are to the line, the stronger the link. A point far from the line is an anomaly.
Dispersion graphs
A dispersion graph plots every value as a dot along a scale, so you can see the spread and any clusters. Putting two sets on parallel lines lets you compare them. In a made-up survey, customers at a corner shop mostly live within 1 km, so their dots bunch together. Customers at a large supermarket live between 0 and 15 km away, so their dots are widely spread. The range is the highest value minus the lowest value.
Describing a trend with data
Use TEA: Trend, Example, Anomaly. This is the same approach as describing a map pattern in the last lesson.
Worked example
A made-up line graph shows visitors to Bay Town, in thousands: 2018 = 40, 2019 = 48, 2020 = 20, 2021 = 35, 2022 = 60. Trend: visitor numbers rose overall. Example: from 40 000 in 2018 to 60 000 in 2022, an increase of 20 000. Anomaly: in 2020 numbers fell sharply to 20 000, which does not fit the rising trend.
Always quote figures with units, and use words such as "rose steadily" or "fell sharply". Describing says what happens. Do not explain why unless the question asks you to.
Common mistakes
Common mistakes
Missing a title, units or axis labels. Uneven scales, for example 0, 10, 20, 50. Starting a bar graph at a number other than zero, which makes small differences look huge. Leaving gaps in a histogram. Bars of different widths in a bar graph. Joining the points on a scatter graph instead of drawing one best-fit line. Pie angles that do not add up to 360°. Describing only one value instead of the whole trend.
Exam-style question
Fig. 1.1 is a made-up table showing the percentage of jobs by type in Town M: farming 10%, factories 30%, services 50%, other 10%. Fig. 1.2 is a made-up table showing visitors to Lake Q, in thousands: 2019 = 30, 2020 = 12, 2021 = 20, 2022 = 32, 2023 = 38.
(a) Calculate the angle for factories on a pie chart of Fig. 1.1. [1]
(b) State two features a pie chart needs so that it can be read. [2]
(c) Describe the trend shown in Fig. 1.2. [3]
(d) Give one annotation that could be added to a line graph of Fig. 1.2. [1]
(e) Evaluate the use of a line graph rather than a bar graph to show the data in Fig. 1.2. [6]
Model answer
(a) 30 x 3.6 = 108° (1)
(b) Any two: a title (1), labels or a key for the slices (1), the values or percentages shown (1).
(c) Trend: visitor numbers rose overall (1). Data: from 30 000 in 2019 to 38 000 in 2023 (1). Anomaly: numbers fell to 12 000 in 2020 (1).
(d) A note on the 2020 point saying it is the lowest value, or a label for the highest value in 2023 (1).
(e) Level 3 (5-6 marks): A line graph is suitable because the data is a series of years, and joining the points shows the trend over time clearly. The fall in 2020 and the steady rise after it are easy to see, and more than one lake could be added with separate lines. A bar graph would make each year easy to compare and would let someone read an exact value from the top of a bar. However it would show the change from year to year less smoothly. A line graph can also look as if it shows values between years, when only five are known. Overall the line graph is the better choice here because the question is about change over time, although a bar graph is better for comparing single years.
Exam tip
When you describe a trend, quote at least two figures from the graph with units, and mention any anomaly.