Week 8 · Data Visualisation with Data Studio I

Three charts are circulating inside Kopi Kembali, and each one is about to mislead a decision-maker. Pick one, take it apart, and rebuild it honestly. This is your Chart Audit and Redesign Brief.

At a glance

Time	60 minutes, alone or in pairs
Work with	Yourself, or a partner
You need	One chart option (rendered chart + CSV) and a spreadsheet, or paper and coloured pencils
You hand in	A redesigned chart + a one-page Chart Audit and Redesign Brief
Graded on	Feedback only — naming the distortion precisely, chart choice, honest axes and labels, one clear message

The scenario

Each option comes as a rendered chart (the artefact you’re critiquing) and the small CSV behind it (the truth you rebuild from). Pick one.

Option	Title	The trap
A	“Skyrocketing Engagement”	A six-week engagement line whose y-axis starts at 45% — the real movement is four percentage points, but the chart shows a cliff face
B	“Revenue correlates with customer growth”	Two lines, two y-axes, moving in perfect sync — check the right-hand scale before you believe the story; what is revenue actually growing on?
C	“Market share”	Six products in a decorated 3D pie — try answering “which product is second, and by how much?” from the pie alone

Get your data

Download the CSV behind your chosen option — your facilitator supplies the rendered chart (PNG) for each option in class.

• Download option-a-engagement.csv — Option A, “Skyrocketing Engagement”
• Download option-b-revenue-customers.csv — Option B, “Revenue correlates with customer growth”
• Download option-c-market-share.csv — Option C, “Market share”

Step 1 — Diagnose · 15 min

Study your chosen chart without opening the CSV. Write down every design decision that shapes what a reader believes: axis start, scales, chart type, ordering, decoration, title. The title is a claim — is it earned?

Step 2 — Check the claim against the data · 10 min

Open the CSV. Compute the honest magnitude of whatever the chart dramatises. State, in one sentence, the gap between what the chart implies and what the data says.

Step 3 — Redesign · 25 min

Rebuild the chart in a spreadsheet, or on paper with real proportions. Apply the unit’s rules:

• Bar lengths and positions start at zero
• Line charts may zoom, but must say so
• One scale per chart, unless there is an overwhelming reason
• Ranking questions get ranked bars
• The title states the honest finding
• Make it work in grayscale

Work through, for your chosen option:

1. What’s wrong with the original? Name the distortion — truncation, dual axes, 3D, area inflation, misleading label, wrong chart type — and explain what it hides or exaggerates.
2. What question should this chart answer? Be specific: “How much does engagement change?”, not “How is engagement?”
3. What chart type answers that honestly? Bar, line, scatter, something else — and why is it better than the original?
4. What changes about the axis, labels, or legend? Start at zero? Remove a line or axis? Reorder categories? Add context or a note?
5. Sketch or build your redesign. Coloured pencils on paper, or a spreadsheet — neat matters less than clear.

Step 4 — Write the brief · 10 min

One page covering: what the original chart made a reader believe, the specific mechanism that did it, what your redesign shows instead, and one sentence stating the correct chart title.

Reflect before you submit

Could the stakeholder misread your new chart? Is there any visual element that carries no information — if so, remove it. If you had to explain this chart to someone who wasn’t in the room, what’s the one sentence you’d say? That sentence is your message — make sure the chart shows it clearly.

Verify & submit

• The dishonest mechanism is named precisely (truncation, dual axes, decoration) — not just “misleading”
• Axis starts at zero (for quantities) and visual area is proportional to the value — double the value, double the length, not four times
• No 3D effects and no dual axes, unless clearly labelled and unavoidable
• Chart type matches the question, and the title is honest
• Legend and labels are clear, and context is stated (vs last month? vs target?)
• Works in grayscale, and uses double-encoding rather than colour alone
• You can defend every element you kept, not only the ones you removed

Submit through the xSiTe checkpoint dropbox folder:

1. A copy of the original chart (photo, screenshot, or sketch)
2. Your redesigned chart (hand-drawn or digital)
3. Your one-page written justification — the distortion, the question the new chart answers, why the design is better (referencing position, length, colour, area), and one sentence on what the reader should understand

This is a portfolio formative checkpoint, feedback only — evidence section 3 of your Individual Analytics Portfolio.

Tip

Feedback covers five things: naming the distortion, chart choice, axis and labels, how well you apply the principle (position, length, colour, area), and whether there’s one clear message.

Companion links: Unit 06 app · lesson plan