Week 9 · Data Visualisation with Data Studio II

Kopi Kembali’s regional manager wants one screen that answers a single question: which outlets and categories should get attention next quarter? Build it as a three-chart interactive dashboard, then have a peer try to break it.

At a glance


Time	80 minutes, one pair per computer
Work with	A partner
You need	Data Studio, plus a spreadsheet or agent for profiling
You hand in	A saved Data Studio dashboard + your Dashboard Design Brief
Graded on	Required individual LO3 evidence — the dashboard, filter, and peer-tested fix all recorded before you leave

The scenario

Kopi Kembali’s regional manager wants one screen that answers: “Which outlets and categories should get attention next quarter?” You have six months of daily sales in kk-daily-sales.csv (about 5,600 rows, January–June 2026).

The dataset is far cleaner than the ones you fought in Units 03 and 04 — but no six-month operational extract is ever perfectly clean. Profile it before you build; anything odd you find belongs in your design brief, and at least one oddity will change how you read a chart.

Data dictionary

Column	Meaning
`date`	Calendar day
`outlet`	One of Kopi Kembali’s outlets
`category`	Kopi & Teh, Specialty drinks, Food, Merchandise
`revenue_sgd`	Revenue for that outlet, category and day
`orders`	Order count for the same

Step 1 — Interrogate before you build · 10 min

In a spreadsheet or with an agent: row count, date coverage per outlet, distinct outlet names across the six months. Write down anything that would distort a time series or an outlet comparison.

Step 2 — Design on paper · 10 min

One question, three charts that each answer part of it, one filter. Decide the trio before touching the tool.

Step 3 — Build · 40 min

Build the three charts, add the one interactive filter, and arrange them so the eye lands on the answer first. Override any default that distorts: date aggregation, sort order, axis ranges, colour.

Chart 1 — bar. Which [category] has the most [metric]? Categories on the x-axis, numbers on the y-axis. Title it with the question it answers, not “Chart 1.”
Chart 2 — line. How does [metric] change over [time period]? Time on the x-axis, in order; no more than three series if you’re comparing lines.
Chart 3 — part-to-whole. What are the slices? A stacked bar is recommended; a pie only if it’s three to four slices.
Filter. One column, linked to all three charts, placed at the top. Test it — when you change it, every chart should update.
Arrange. Most important chart first, or centre-right; a title at the very top; leave white space. Run the squint test: do the critical numbers stand out, or are you distracted by gridlines and noise?

Step 4 — Peer-test, fix, and write the brief · 20 min

Exchange dashboards with a peer. They test the filter, the labels, the data freshness, and check at least one chart value against the source CSV. Fix one issue they find, and record the change.

Then write the brief. For each chart: the sub-question it answers, the chart type and why, and one default you overrode and why. Add the caveats a reader needs — this is where your Step 1 findings pay their way. A dashboard that presents a distorted week without a footnote is a distorted dashboard.

Verify & submit

The dashboard answers its one question without the viewer hunting
Every chart type matches its sub-question, and you can say why
The filter changes all three charts sensibly
Date coverage and naming oddities are handled or footnoted, not silently plotted
The brief records at least one overridden default per chart
One chart value is checked against the source data
The peer test is recorded, with the fix you made

Share the dashboard (view-only link) and add the link plus one evidence screenshot to your individual portfolio, along with this worksheet. Submit the Unit 7 xSiTe checkpoint by the end of the session — required individual LO3 evidence for your Individual Analytics Portfolio.

Capstone connection

Your three-chart dashboard this week is a template. The capstone dashboard will be larger and more complex, but the thinking is identical: start with the stakeholder’s question (not “what data do we have?”), choose charts that answer it one insight at a time, add filters that let the stakeholder explore, and arrange them so the story is clear in ten seconds. In your capstone, your team will connect a Google Sheets data source directly to Data Studio, so the cleaned data from Stages 1 and 2 flows straight into the dashboard — the mechanics of that connection are in Tool Module B.

Companion links: Unit 07 app · lesson plan