← BIT 304 Data Science and Analytics
Data Visualisation and Summarising Findings
Notes
Matching charts to questions
Data visualisation represents information graphically to reveal comparisons, patterns and uncertainty. Bar charts compare categories, line charts show ordered changes over time, histograms display numerical distributions and scatter plots examine relationships between numerical variables. Choose the chart according to the analytical question, not visual novelty. Label axes, units and time periods clearly, and use colours that remain distinguishable for readers with colour vision differences. For bar charts, a zero baseline generally preserves honest comparisons of length. A chart should make the evidence easier to interpret rather than exaggerate small differences.
Turning evidence into a summary
A useful analytical summary states the question, dataset, main finding, limitation and recommended action. Counts show scale, while percentages allow comparisons across groups of different sizes. A percentage point change is the difference between percentages; percentage change divides the difference by the original value. Correlation describes association, not proof that one variable causes another. Report sample sizes and relevant uncertainty, particularly when groups are small. Avoid selecting only favourable periods or categories. Tables complement charts when exact values matter, while concise annotations can explain events that affect interpretation.
Practical example
A Mombasa wholesale shop compares deliveries across two months. In April, 72 of 90 deliveries arrived on time, producing 72 divided by 90 times 100, or 80 percent. In May, 102 of 120 arrived on time, giving 85 percent. The increase is five percentage points, while relative improvement is 5 divided by 80 times 100, or 6.25 percent. A labelled bar chart can compare the rates, with delivery counts shown alongside. The manager should not claim that new routing software caused the improvement without examining distance, staffing and order mix, which may also have changed between months.
Quiz
1. An on-time rate rises from 80 percent to 85 percent. What is the percentage point increase?
2. Why should delivery counts accompany rates?
3. Which chart is most suitable for monthly sales over two years?
Sign in to submit and securely grade your quiz.