Sorted descending bar, title like "Furniture leads revenue at $24.8M —
five to one over Decor." What alphabetization hid: the ranking itself. Alphabetical order
forces the reader to hop bars mentally reconstructing what a sort would have simply shown;
the two largest categories sat in positions two and five.
Line chart of [Total Revenue] by month across all years — the growth
drift and repeating seasonal humps appear immediately as slope and rhythm, invisible in the
42-column forest. With [Revenue LY] added: two lines whose vertical gap is YoY
growth, readable month by month. Title: "Every 2026 month runs ahead of its 2025
counterpart." The species change did what no formatting of columns could.
Sorted bar of the top subcategories (Top N filter, or all 20 sorted —
scrollable bars beat unreadable angles either way), title like "Seating and Tables are
the subcategory heavyweights." The 20-slice pie could not be ranked by eye past the
first slice or two; angles within a few degrees are indistinguishable.
Reflection answers
(1) The alphabetical bar answered "what is the revenue of each category, in dictionary
order?" — a lookup nobody asked for. (2) The 42 columns answered "how does each individual
month compare to every other month?" — comparison, when the question was movement. (3) The
pie answered "do twenty parts make a whole?" — technically yes, uselessly. Wrong species =
right answer to the wrong question.
The makeover consisted entirely of choosing the claim: species, sort,
and title. Zero new DAX, zero new data. Module 2 built what the numbers are; this module is
about what the numbers say — and the same measures said almost nothing in the before-charts
and said it plainly in the afters.