Data Visualisation and Interpretation

A practical guide to presenting survey results clearly, choosing suitable charts and interpreting patterns without overstating what the data can show.

AC 2.2 moves from calculation to visualisation and interpretation. Using the survey results calculated in AC 2.1, you need to present findings through two different chart formats: one showing mean scores and another showing survey responses as percentages. You then need to interpret what the visualised data shows.

This is a practical criterion. A successful response needs accurate charts, appropriate chart selection, clear labels and meaningful interpretation. This guide therefore uses illustrative examples rather than creating charts from the Camellia assessment dataset.

A table can contain accurate information but still make patterns difficult to recognise quickly. Data visualisation converts numerical results into a form that can make comparisons, concentrations, differences and unusual results easier to identify.

The purpose of a chart is not decoration. It should help the reader understand an important feature of the evidence more efficiently than the raw table alone. A visually impressive chart that makes comparison difficult is less useful than a simple chart that communicates the pattern clearly.

Percentage data contains several response categories for each statement. A 100% stacked bar chart can be useful because each bar represents the complete response distribution and the segments show the proportion selecting each category.

This allows the reader to see whether responses are mainly positive, mainly negative, concentrated around the middle or divided between different views. Other chart formats may also be suitable, but the chosen format should make comparisons across statements clear.

Imagine a practice statement with the following distribution: 10% Strongly Disagree, 15% Disagree, 20% Neutral, 30% Agree and 25% Strongly Agree. A percentage chart would show not only that positive responses are relatively common but also that a substantial minority selected neutral or negative categories.

This is information that can be obscured by a single mean score. The example is illustrative only and is not drawn from Camellia’s survey data.

A mean summarises the response distribution into one value, which makes comparisons across statements convenient. However, different distributions can produce similar means.

For example, a statement where most employees select the middle category could have a similar average to one where employees are sharply divided between positive and negative responses. The percentage chart exposes that distribution.

Using both forms of visualisation therefore gives a fuller view of the survey than relying on either one alone.

Every chart should be understandable without requiring the reader to guess what the visual represents. Use a clear title and label the relevant axis or categories. Where colours or patterns represent response categories, provide a clear legend.

Statement labels should remain readable. If the original survey wording is too long for an axis, use concise identifiers or shortened labels and make clear what they refer to elsewhere in the document.

Data labels can be useful where exact values matter, but adding labels to every possible element can make a chart crowded. Include information that helps interpretation rather than displaying everything simply because the software allows it.

Consistency makes a set of charts easier to understand. Response categories should appear in a logical order and should not unexpectedly change position between charts. Formatting conventions, decimal places and terminology should also remain consistent.

Avoid unnecessary three-dimensional effects, excessive decoration or visual elements that compete with the data. For assessment work, clarity and accuracy are more important than elaborate styling.

Neutral responses can be analytically important. A large neutral proportion might reflect genuine neutrality, uncertainty, mixed experiences or a lack of sufficient information. The survey alone may not establish which explanation is correct.

Similarly, a divided distribution can disappear inside an average. If many respondents select opposite ends of the scale, the mean may sit near the centre even though few people actually hold a middle view. This is another reason to interpret percentage distributions alongside means.

A survey can show how respondents answered particular questions, but it does not necessarily establish why those responses occurred. If employees report less positive views about development, for example, the chart itself does not prove whether the cause is access, quality, communication, management support or another factor.

Interpret what the data supports and identify areas for further investigation where necessary. This is more credible than inventing an explanation.

For Camellia, the calculations from AC 2.1 provide the basis for the two visualisations. The mean-score chart should help compare overall results across survey statements, while the percentage chart should reveal how responses are distributed.

Once the charts are created, identify the most important patterns and explain what they indicate about the survey results. Those findings can then provide an evidence base for the recommendations considered in AC 2.3.

Do not decide the interpretation before examining the completed charts. Let the calculated evidence determine which findings deserve attention.

• Does the chart use the correct AC 2.1 calculations?
• Is the chosen chart appropriate for the type of data?
• Is there a clear and informative title?
• Are axes, response categories and legends labelled correctly?
• Are percentage categories presented in a logical and consistent order?
• Is the numerical scale potentially exaggerating or hiding differences?
• Can long survey-statement labels still be understood?
• Do displayed values use consistent rounding?
• Have you checked that the chart range includes all intended spreadsheet cells?
• Does the interpretation match what the chart actually shows?

• Choosing a chart because of appearance rather than analytical purpose.
• Using the wrong data range when creating the chart.
• Leaving charts without meaningful titles, labels or legends.
• Using a pie chart for many statements or categories where comparison becomes difficult.
• Using 3D effects that distort visual comparison.
• Repeating every number instead of interpreting significant patterns.
• Discussing mean scores without examining the response distribution.
• Claiming that the survey proves why employees hold particular views.
• Making recommendations before clearly identifying what the data indicates.