Your starting point
Prerequisites: Use the ideas and vocabulary from lesson 18; revisit its worked examples if needed.
Definition: A chart maps data values to visual properties such as position, length or area. The appropriate chart depends on whether the question concerns category comparison, change over ordered time, a distribution, paired association or parts of a whole.
Start with the comparison
Write the question before choosing the display. “Which channel had more sessions?” suggests category comparison. “How did sessions change over time?” requires an ordered sequence. “Do longer visits tend to have higher spending?” needs paired observations. Selecting a fashionable chart first can force the data into a visual structure that implies relationships you did not measure.
Match the marks to the evidence
Bars support comparisons of magnitude across categories. Lines emphasize ordered continuity. Scatterplots retain pairs of numerical values. Histograms group a numerical distribution into intervals. Pie charts can describe a small number of mutually exclusive positive parts of one total, but are often less precise for comparison than bars. A collection of percentages that overlap is not a valid pie merely because software can draw it.
Make the display self-contained
Supply a meaningful title, labels, units, time period and a statement of any filtering or aggregation. Preserve an accessible table of plotted values and do not encode essential information by color alone. A reader should be able to distinguish an individual observation from a group mean. Export the chart only after comparing its values with the underlying summary. A beautiful chart with an ambiguous denominator is still a poor explanation.
Two worked examples
Channel comparison
Search has 120 sessions, Email 80 and Social 40 in a synthetic period. A bar chart compares the three unordered categories. Connecting them with a line would imply an order or transition that the channel labels do not provide. If you reorder bars by value, say that the order is a presentation choice.
A part-to-whole question
A budget contains mutually exclusive amounts 50, 30 and 20, totaling 100. A pie can represent those shares. If the categories instead represent overlapping survey interests, adding them may exceed the number of respondents. Use separate percentages or bars rather than pretending overlapping answers form disjoint slices.
Where this becomes useful
A workshop facilitator presents learning participation to a mixed audience. One simple chart with clearly labeled counts and a short limitation is often easier to inspect than several decorative visuals with inconsistent definitions.
Make an attempt
Complete Chart Selection Laboratory. For each question, choose a chart and explain the variable types before reading the feedback. Then create one chart from a synthetic dataset and inspect its exact plotted-value table.
Misconception check
Claim: More visual complexity makes an analysis more advanced.
Inspect the correction
Complexity is useful only when it helps answer the question. Simpler encodings often make errors and comparisons easier to inspect.
Independent application
Take one synthetic dataset and write three different questions. Choose a suitable chart for each, explain why another chart would mislead, and create one labeled display with an accessible value table and a limitation statement.
Completion records practice and self-review. It is not certification of statistical competence.
