Your starting point
Prerequisites: Use the ideas and vocabulary from lesson 19; revisit its worked examples if needed.
Definition: A bar chart represents numerical magnitude through bar length across discrete categories. Because length is the comparison cue, a conventional bar chart uses a zero baseline. Aggregation determines whether each bar represents a sum, mean, count or another summary.
Define what one bar means
A bar labeled North might show total revenue, average order revenue or number of orders. These are different summaries of the same records. State the aggregation and unit explicitly. A larger total can result from more observations rather than a higher typical value. Include counts when unequal group sizes matter to the question.
Use ordering and labels deliberately
Sort bars by value when rank is the question, or use a natural order when categories have one. Avoid implying a meaningful order where none exists. Long labels need readable placement rather than tiny text. Too many categories can make the chart unusable; filter with a stated rule or create a carefully documented “other” group rather than silently dropping inconvenient values.
Preserve proportional lengths
Starting a bar axis far above zero exaggerates relative differences in length. If small changes need emphasis, a point-based display with explicit scale can be more appropriate than truncated bars. Negative values require an axis that shows the zero crossing. The workbench includes zero and supplies a table of plotted values so the visual can be checked independently of its appearance.
Two worked examples
Total versus mean
North has two orders of 50 each; South has one order of 80. Total revenue bars are North 100 and South 80. Mean order revenue bars are North 50 and South 80. Neither chart is wrong if clearly labeled, but they support different conclusions about volume and typical order size.
A baseline problem
Values 100 and 110 differ by ten units, or 10% relative to 100. Bars beginning at 95 have visible heights 5 and 15, making the second appear three times as tall. Bars beginning at zero preserve the intended magnitude comparison. Axis labels are necessary, but a misleading length cue can still dominate perception.
Where this becomes useful
A team compares requests by support channel. A count chart can guide staffing, while a mean-duration chart can guide process investigation. Displaying the two summaries separately is clearer than calling both “channel performance.”
Make an attempt
Load Retail sales in the workbench. Set region as the category and revenue as the value. Compare Sum and Mean, inspect the plotted-value table, and explain what changed in the question. Verify that zero remains in the scale.
Misconception check
Claim: The tallest total bar identifies the group with the best average outcome.
Inspect the correction
A total combines group size and individual values. Inspect means, counts and context before judging typical performance.
Independent application
Create two bar charts from the same grouped data: one showing totals and one showing means. Include units, counts and the filter scope. Write a separate finding for each and explain why a decision-maker should not interchange them.
Completion records practice and self-review. It is not certification of statistical competence.
