Data Visualisation
A chart is an argument about what matters. Choose the form that makes the comparison honest and the takeaway obvious.
Success criteria
- I can choose a chart type that fits my data and purpose.
- I can label a chart so it can be read without guessing.
- I can recognise common distortions in graphs.
- I can annotate the takeaway instead of leaving the reader to find it.
Choose the right form
| If the data shows... | Use... | Avoid... |
|---|---|---|
| Categories compared by size | Horizontal or vertical bar chart | A pie chart with many slices. |
| Change over time | Line chart | A bar chart with uneven time gaps. |
| Parts of a whole | Stacked bar or a small pie with few slices | 3D pie charts. |
| One key number | A large figure with a short context line | A chart that adds no comparison. |
| Ranking | Sorted bar chart | Alphabetical order that hides the pattern. |
| Location | Map with a clear legend | A map with no scale or legend. |
| A process | Flowchart, cycle, or timeline | A line chart pretending the steps are numbers. |
| Relationships | Diagram with labelled connections | Arrows with no explanation. |
| Uneven quantities | Area or treemap, used carefully | Comparing circles by diameter. |
Anatomy of an honest chart
Model chart
This chart uses illustrative values to demonstrate layout and annotation. It is not real data.
Recycling rate by material (illustrative)
Annotation: paper is recycled at the highest rate; soft plastic is the clear outlier and needs a separate solution.
Source: illustrative only. Replace with a real dataset and source line.
Annotation does the thinking
Do not make the reader interpret the chart alone. Add a short annotation that states what the reader should notice.
Weak
"This graph shows recycling rates."
Stronger
"Paper is recycled at more than six times the rate of soft plastic, so the biggest gain would come from a separate soft-plastic collection."
How graphs mislead
| Trick | What it does | What to check |
|---|---|---|
| Truncated axis | Makes a small change look huge. | Does the vertical axis start at zero? If not, is the break marked? |
| Inconsistent intervals | Stretches some years and squeezes others. | Are the time gaps equal? |
| Area distortion | Doubles the width and height, so the area looks four times larger. | Is size proportional to the value? |
| 3D effects | Changes the apparent size of bars or slices. | Can you read the values without the perspective? |
| Too many pie slices | Makes small differences impossible to judge. | Would a sorted bar chart be clearer? |
| Cherry-picked range | Hides a longer trend by choosing a convenient start date. | What happens if the time range is extended? |
| Missing denominator | "Doubled" could mean 2 to 4, or 2,000 to 4,000. | What are the actual values? |
| Correlation as cause | Implies one trend caused another. | Is there a mechanism and independent evidence? |
Accessible charts
- Use direct labels rather than a colour key.
- Add patterns or text labels if colour carries meaning.
- Write a one-sentence summary of the chart for readers who cannot see it.
- Keep the chart uncluttered. Remove gridlines that do not help comparison.
- Use a font size that survives being viewed on a phone.
Chart planner
| Question | My answer |
|---|---|
| What comparison or change am I showing? | |
| What chart type fits best? | |
| What is the unit and scale? | |
| What is the source and year? | |
| What should the reader notice? | |
| How will I label it without relying on colour alone? |
If you do not have reliable numbers, do not invent a chart. Use a labelled diagram, a process, or a comparison of evidence instead. A clear diagram with a real source is stronger than a fake graph.
Assessment link
Data visualisation supports the Information and ideas and Visual language and design criteria. In your rationale, explain why the chart type fits the data and how the annotation guides the audience.