Honest charts and claims that stay in the data

A bar chart of your three supplies should start its value axis at zero, label its units, and total to the same 29 as the table. A truncated axis makes a small difference look dramatic — which is occasionally a legitimate choice and far more often an accident you'd never have made deliberately.

Check the chart against the source table rather than against your impression of it. Bars in the right order, values matching, nothing rounded into a different story. If chart generation isn't available in your account, build it in a spreadsheet from the verified table — the checking discipline is the lesson, not the tool.

Then the summary, which is where claims escape the data. "Boxes contribute $15 of the $29 total" is in the dataset. "Boxes cost the most because people buy more storage than furniture" is not — there is no customer, no behaviour and no time series in three rows of supplies. It's a plausible sentence describing a world the data says nothing about.

Ask for explanations bounded by the source: which item contributes most, and by how much. If a cause is offered, it should be marked as a hypothesis to test, not delivered as a finding. This is the same instinct as a source-bound answer in Level 2 — the boundary just happens to be a table this time.

Loading your workspace…