Did you know that a pupil can make a beautiful graph that leaves another pupil unable to discover its main finding? A chart can be technically correct and still be inaccessible when its meaning exists only in colours, bar heights or tiny labels. The good news is that explaining a chart in words can deepen the author’s own understanding, too.
A text alternative for a graph explains its purpose and essential information to a person who cannot rely on the image. For a simple image, a concise description may suffice. For a complex chart, use a brief alternative identifying the chart and its finding, plus a nearby detailed description or accessible data table. The immediate answer is to state what is plotted, give the key values and comparison in words, and preserve access to all data needed for the reader’s task—without inventing an explanation for the pattern.
Start by asking what the chart lets a sighted reader know
Suppose a science report shows four bars labelled Monday to Thursday, and a learner’s only description is ‘colourful bar chart’. The words identify the type of graphic but not the question it answers. They do not tell a screen-reader user the units, what was measured, or which days differed. A useful diagnostic is to cover the picture and read only the accompanying text. Can the listener identify the highest value, the lowest value and the conclusion supported by the data?
The W3C Web Accessibility Initiative guidance on complex images explains why charts often need a short alternative identifying the image and a more complete representation of the important information. The guidance gives examples of data tables and descriptions for charts. It does not say that every chart must receive one fixed-length sentence. The amount of description depends on the image’s purpose and the information readers need.
A hand-checked example: daily minutes of independent reading
The following invented figures are illustrative, not observations of real students. A class activity records total independent reading minutes by one fictional pupil over four school days: Monday 20 minutes, Tuesday 25 minutes, Wednesday 18 minutes, Thursday 30 minutes. A bar graph uses days on the horizontal axis and minutes on the vertical axis. Assume all readings were recorded using the same rule and represent minutes rather than number of books.
The shortest adequate summary depends on the chart’s purpose. If the assignment asks which day had the most reading, one helpful text alternative could be: ‘Bar chart of a pupil’s reading minutes from Monday to Thursday; Thursday is highest at 30 minutes and Wednesday lowest at 18.’ A nearby description could give all four values in order so that no information is reserved for a sighted reader.
The accessible data behind the picture
- Monday — 20 minutes of reading.
- Tuesday — 25 minutes of reading.
- Wednesday — 18 minutes of reading.
- Thursday — 30 minutes of reading.
Check the arithmetic. Total recorded time is 20 + 25 + 18 + 30 = 93 minutes. The range is 30 − 18 = 12 minutes. The mean over the four recorded days is 93 ÷ 4 = 23.25 minutes per day. If the learner says ‘the pupil read more every day’, the claim is false: Tuesday’s 25 minutes fell to Wednesday’s 18. If the learner says ‘Thursday had the most recorded reading’, the chart supports it.
A good full description might read: ‘The horizontal axis lists Monday, Tuesday, Wednesday and Thursday. The vertical axis shows reading minutes. The bars represent 20, 25, 18 and 30 minutes respectively. Thursday has the largest recorded value and Wednesday the smallest; the sequence does not rise steadily.’ These sentences let someone reconstruct the comparison without access to colour. Because the values are invented, they cannot establish any universal finding about reading habits.
Three layers of information—and why each layer matters
- Identity: What is this image? For example, ‘bar chart of reading minutes over four days.’ Avoid a vague ‘graph’ or repeating a heading that already says exactly the same thing.
- Takeaway: Which comparison matters for this question? For example, Thursday is highest, Wednesday lowest and the series is not steadily increasing.
- Recoverable data: Provide exact categories, units and values in adjacent accessible text or an appropriately structured data table when readers may need to make their own calculations.
Do not confuse a chart description with a fully interpreted essay. A statement such as ‘Wednesday was lower because the pupil became lazy’ is not information present in the graph. The chart contains minutes, not motivation. Even ‘Thursday was best’ depends on the purpose: more minutes may mean greater engagement, but a single tally does not measure comprehension, enjoyment or the quality of reading.
What changes for other types of school graphs?
Line graph: preserve sequence as well as endpoints
If a line graph shows a temperature changing from 20 °C to 25 °C, then 18 °C, then 30 °C over four observations, a description that gives only the first and last readings hides the dip. Name the ordered readings, measurement times and units. Do not call a curved line ‘exponential’ without evidence or a stated model.
Grouped bar chart: identify which series goes with which category
Imagine two classes each have a bar for every subject. Saying ‘the blue bars are taller’ is unreliable without indicating which series is blue. Name the categories and series directly. This extends the checking method in our related article, how category order changes grouped-bar-chart reading. When groups have different totals, consider whether raw counts and percentages answer different questions.
Pie chart: identify the whole and how values were produced
A pie chart description should identify each category’s share and what 100% represents. If the sample comprises only 20 replies, a 50% slice means 10 of the 20 replies under a single-response rule; it does not automatically mean 50% of the entire school. State when counts are estimates or percentages have been rounded.
A progression from guided description to independent communication
- Guided observation: have the teacher point out the title, axes, units and largest and smallest values, then write a sentence together.
- Supported reconstruction: hide the graph and read the learner’s draft description aloud. Ask a partner to sketch rough bars or restate the ordering from the text alone.
- Independent description: give the learner a fresh graph and ask for a short alternative plus a data-rich description without teacher prompts.
- Evidence check: compare every word of the description with the plotted measurements. Cross out causal claims or universal statements that the data do not support.
- Transfer: repeat the task with a line graph, grouped bar chart or pie chart. Explain what additional information the new graph requires.
What parents, teachers and pupils can actually observe
A parent need not know any screen-reader software to start. Ask the child to explain a graph over the phone as though the listener cannot see it. Can the listener answer the same question as someone looking at the graphic? A teacher can assess whether units are reported, key comparisons are correct, all necessary values are recoverable and unsupported claims are excluded.
For school websites and digital assignments, provide descriptions in the actual publishing tool and check that they are exposed to assistive technology. Do not simply place a tiny caption inside the picture. Testing with a screen reader or a reader who uses accessible text can reveal shortcomings that a visual inspection misses. The wider W3C images tutorial distinguishes decorative, informative and complex images; choosing the right alternative is part of the author’s responsibility.
Limits and easy misconceptions
There is no magic alt-text word count. Some simple figures need only a brief alternative, while detailed school data may require a well-structured table or description outside the alternative-text field. A full table of data may be clearer than a 300-word string crammed into an image attribute. Another limitation is uncertainty: a photograph of a blurry hand-drawn graph may not let you read precise values. Say ‘approximately’ or request the underlying data rather than presenting guesses as exact measurements.
Another common mistake is to use only colour: ‘red goes up, green goes down’. Some readers do not see colour differences reliably. Identify each series by name and describe its pattern explicitly. Accessibility does not mean oversimplifying the numbers; it means making the same evidence available through another form.
Frequently asked questions
Should the description repeat every number on the chart?
When readers need to calculate, compare or verify the data, provide all essential values, often best as a nearby accessible table or ordered list. When the image is only a small illustrative icon and the key fact is already stated in the surrounding text, a shorter alternative may be appropriate.
Can the description say what caused a rise or fall?
Only if the graphic and accompanying evidence establish that explanation. A graph of reading minutes does not reveal why Wednesday was lower. Separate ‘the graph shows’ from ‘one possible explanation is’.
Does a graph become accessible just because it has a title?
No. A title names the subject but may omit axes, units, comparisons and actual values. Test what a reader can understand without seeing the chart.
Can an accessible description help pupils who can see the image?
Yes. A carefully written description forces the author to check labels, units and interpretations. It can also aid readers on a slow connection or people who find dense visual displays difficult to parse.
The point of the exercise
The eduKate Education Hub brings together ways to make knowledge independently usable, and our immutable Clementi Secondary 1 Mathematics tutorial reference illustrates diagnosis, explanation and checking in a practical learning sequence.
The core aim is not to fill an accessibility field for its own sake. It is to help another person recover the same evidence, make the same defensible comparisons and question the same limitations. When a pupil can describe a graph faithfully without pointing at it, the pupil has often learned to read it more accurately as well.