T3 lab: One page for the leadership team¶

Rebuild the Session 14 decision dashboard in Tableau: KPI tiles, a monthly trend and a region chart that filters the rest of the page when you click it. Check that every number still reconciles after a click, write a message headline, and make sure the page works without colour.

This lab rebuilds the decision dashboard from the Session 14 lab (task I1, retail case) in Tableau. The Stretch task builds a second dashboard on the marketing case from the Session 12 lab. If you get stuck, look back at how you built the same page in Python.

---

The business decision¶

The commercial lead from Session 14 is back:

"Next week I present our plans to the leadership team. I need one page that shows how 2025 went against 2024, and whether anything should worry us before we expand. They will want to click a region and see its numbers. Make it impossible to misunderstand."

The decision is where to focus next year and what to check before expanding. The page must answer it in a minute, and it must stay correct when someone clicks.

---

How this lab works¶

Part	What you do	Tasks
Setup	Open your T2 workbook, check its fields and add the year fields, everyone together	Setup
Core	Build the tiles, trend and region chart; lay out the dashboard; add a filter action and reconcile; finish, check and export	C1–C4
Stretch	Build a one-page dashboard for the marketing case, with a filter action and a highlight action	S1
Challenge	Critique a deliberately poor dashboard, then prove two of its problems in Tableau	X1–X2

Each task has a level:

• Core: everyone completes it; it is enough for your portfolio.
• Stretch: for when you finish the core tasks early.
• Challenge: optional, harder reasoning.

---

What you will be able to do¶

1. Build KPI tiles that show a value for one year with a comparison.
2. Lay out a dashboard in reading order with containers and a fixed size.
3. Add a filter action and check that the tiles still reconcile for every selection.
4. Write a message headline and a scope note that stay true when the page is filtered.
5. Check a dashboard for accessibility, including without colour.
6. Add a phone layout and export the dashboard as a PDF.

Evidence you keep: your saved workbook, your dashboard PDF, your region reconciliation table, your headline and accessibility check, and any Stretch and Challenge notes.

Before you start, work through the T3 lecture. Have your T2 workbook, your evidence log and your Session 14 dashboard image to hand.

---

Setup: open your T2 workbook¶

Step 1: Save a copy¶

1. Open Tableau Desktop Public Edition and open T2_lab_YourName.twbx.
2. Choose File → Save As and save it as a Tableau Packaged Workbook (*.twbx) named T3_lab_YourName.

Step 2: Check the data source¶

1. Click the Data Source tab at the bottom left.
2. Check that Orders and Returns are on the canvas with a relationship line between them.
3. Check that the Filters area at the top right shows 1 filter: Order Date from 1 January 2023 to 31 December 2025.

Step 3: Check your fields¶

Go to any sheet and look in the Data pane. You need these fields from T1 and T2, with exactly these names:

Field	Formula	Made in
Returned Order	NOT ISNULL([Order ID (Returns)])	T1 setup
Distinct Orders	COUNTD([Order ID])	T1 C1
Profit Ratio	SUM([Profit]) / SUM([Sales])	T2 C1

Create any field that is missing with Analysis → Create Calculated Field....

Step 4: Add the year fields¶

The tiles show 2025 with a comparison to 2024, so they need both years in the same sheet (T3 lecture, Part 3). Create these calculated fields. Check that each says The calculation is valid before you click OK.

Field	Formula	Format
Sales 2025	SUM(IF YEAR([Order Date]) = 2025 THEN [Sales] END)	Number, 0 decimals, thousands separator
Sales 2024	SUM(IF YEAR([Order Date]) = 2024 THEN [Sales] END)	Number, 0 decimals, thousands separator
Sales Growth	[Sales 2025] / [Sales 2024] - 1	Percentage, 1 decimal
Orders 2025	COUNTD(IF YEAR([Order Date]) = 2025 THEN [Order ID] END)	Number, 0 decimals, thousands separator
Profit Ratio 2025	SUM(IF YEAR([Order Date]) = 2025 THEN [Profit] END) / [Sales 2025]	Percentage, 1 decimal
Returned Orders 2025	COUNTD(IF YEAR([Order Date]) = 2025 AND [Returned Order] THEN [Order ID] END)	Number, 0 decimals
Returned Share 2025	[Returned Orders 2025] / [Orders 2025]	Percentage, 1 decimal

To set a format: in the Data pane, right-click the field and choose Default Properties → Number Format.... For a number with no decimals, choose Number (Custom), set Decimal places to 0 and tick Include thousands separators.

In plain words:

• Sales 2025 keeps the Sales of order lines placed in 2025, then adds them up.
• Orders 2025 counts each 2025 order once. Lines from other years give a blank, and COUNTD ignores blanks.
• Returned Share 2025 is a ratio of counts: orders on the returns list divided by all orders, both for 2025. It is not an average of anything.

Save your workbook (File → Save).

---

Core tasks¶

C1 (Core): Build the sheets¶

You need six sheets: four tiles, one trend and one region chart. Build each on its own sheet; the dashboard comes in C2.

Part A: The first KPI tile¶

1. Create a new sheet and name it Tile Sales.
2. Drag Sales 2025 onto Text on the Marks card. The view shows one number.
3. Drag Sales Growth onto Text as well. Both pills now sit on the Marks card.
4. Click Text on the Marks card, then click the ... button. The Edit Label window opens.

5. Delete what is there and write three lines. Use Insert to add the fields:

   Sales, 2025
   <AGG(Sales 2025)>
   <AGG(Sales Growth)> vs 2024
   

6. Select the middle line and set a large font size, for example 24, in bold. Set the other two lines to about 10. Click OK.

7. Right-click the sheet title and choose Hide Title. The first line of the tile is its label.

Part B: Three more tiles¶

Duplicate Tile Sales three times (right-click the tab → Duplicate). In each copy, drag the old pills off the Marks card, drag the new fields onto Text, and edit the text:

Sheet	Fields on Text	Label text
Tile Orders	Orders 2025	Orders, 2025 / value / distinct orders
Tile Profit Ratio	Profit Ratio 2025	Profit ratio, 2025 / value / Profit ÷ Sales
Tile Returns	Returned Share 2025, Returned Orders 2025, Orders 2025	Orders on returns list, 2025 / share / <Returned Orders 2025> of <Orders 2025> orders

Part C: The trend¶

The Session 14 trend drew 2025 against 2024, month by month, so the reader can see whether growth was steady or came from one lucky month.

1. Create a new sheet named Trend.
2. Drag Order Date to Columns. Right-click the pill and choose the first Month in the menu (the blue one, with an example such as May). The axis now shows the twelve months of the year.
3. Drag Sales to Rows.
4. Drag Order Date onto Color on the Marks card. Tableau shows YEAR(Order Date): one line per year.
5. Drag Order Date to the Filters shelf, choose Years, click Next, tick 2024 and 2025 only, and click OK.
6. Set the colours: click Color → Edit Colors.... Give 2024 a grey and 2025 a blue. Click OK.
7. Make the years readable without colour: drag Order Date onto Label (it shows YEAR(Order Date)). Click Label, tick Show mark labels, and under Marks to Label choose Line Ends.
8. Check the vertical axis starts at zero: right-click it, choose Edit Axis..., and make sure Include zero is ticked.
9. Title the sheet, for now descriptively: Monthly Sales, 2025 against 2024 (file's own monetary units). You rewrite it as a message title in C4.

Part D: The region chart¶

1. Create a new sheet named Returns by region.
2. Drag Region to Rows and Returned Share 2025 to Columns.
3. Click the Sort descending icon in the toolbar, so the longest bar is at the top.
4. Drag Returned Share 2025 onto Label, so each bar shows its value.
5. Drag Returned Orders 2025 and Orders 2025 onto Tooltip. Hover over a bar to check both appear.
6. Drag Region onto Color. Click Color → Edit Colors.... Give every region the same grey, then give the one region your message is about an orange. Click OK.
7. Title the sheet: Share of 2025 orders on the returns list, by region.

Look at the chart and decide which region your message is about. Write one sentence in your log saying why.

Save your workbook.

C2 (Core): Lay out the dashboard¶

Before you start, sketch the page on paper: headline, a row of four tiles, the trend across the full width, the region chart, and the scope note at the bottom.

1. Choose Dashboard → New Dashboard. Name the tab Decision dashboard.
2. In the Dashboard pane on the left, under Size, choose Fixed size and Generic Desktop (or another fixed size your facilitator gives). Do this first.
3. Under Objects, check that Tiled is selected at the bottom.
4. Drag a Horizontal object from Objects to the top of the page. This container holds the tiles.
5. Drag the four tile sheets, one by one, into the horizontal container. Watch for the grey shaded area: it shows where the sheet will land. Order: Sales, Orders, Profit Ratio, Returns.
6. Click the container's edge to select it. Open its drop-down menu (the small arrow at the top right) and choose Distribute Contents Evenly.
7. Drag Trend below the tiles, across the full width.
8. Drag Returns by region below the trend.

9. Drag a Text object to the bottom of the page and write the scope note:

   Scope: Superstore orders placed in 2024 and 2025 (fictitious sample published by Tableau).
   Sales in the file's own monetary units. Profit ratio = Profit ÷ Sales. Returns list = which
   orders appear on it; no dates, amounts or reasons. Click a region to filter the tiles and trend;
   click it again to clear. Patterns describe what happened, not why.
   

10. Choose Dashboard → Show Title. Leave the default title for now; you write the headline in C4.

Save your workbook.

C3 (Core): Add a filter action, then reconcile¶

The leadership team wants to click a region and see its numbers.

Part A: Add the action¶

1. Choose Dashboard → Actions... and click Add Action → Filter....
2. Name the action Filter by region.
3. Under Source Sheets, choose Decision dashboard in the drop-down and tick only Returns by region.
4. Under Run action on, choose Select.
5. Under Target Sheets, tick the four tile sheets and Trend. Untick Returns by region, so the reader can still see and click every region.
6. Under Clearing the selection will, choose Show all values.
7. Click OK, then OK again.

Click one region's bar on the dashboard. The tiles and the trend now show that region only. Click the same bar again to clear the selection.

Part B: Reconcile across regions¶

A tile that changes when you click must still add up. Click each region in turn and copy the tile values into a table in your log:

Region	Sales 2025	Orders 2025	Returned Orders 2025	Returned Share 2025
(first region)
(second region)
(third region)
(fourth region)
Added up				(do not add)
No region selected

Then check, and write each answer in your log:

• Do the four regions' Sales 2025 add up to the value with no region selected? Why must they?
• Do the four regions' Orders 2025 add up to the value with no region selected? If they do, what does that tell you about orders and regions? Compare with categories in T2 C1 Part C.
• Do the four regions' Returned Orders 2025 add up?
• Divide the added-up Returned Orders by the added-up Orders with a calculator. Does it equal the Returned Share 2025 tile with no region selected?
• Now take the simple average of the four regions' shares. Is it the same? Which one is the company's share, and why? Hint: T2 lecture, Part 2.

Part C: Reconcile the tiles with the trend¶

1. Right-click the Trend sheet tab and choose Duplicate as Crosstab. Tableau creates a table of Sales by month and year.
2. Choose Analysis → Totals → Show Column Grand Totals (if the totals appear across instead of down, try Show Row Grand Totals).
3. The 2025 total must equal the Sales 2025 tile with no region selected. Divide the 2025 total by the 2024 total and subtract 1. The result must equal the Sales Growth on the tile.
4. Count the months in which 2025 Sales are higher than in the same month of 2024. Write the count in your log; you use it in your headline.

Save your workbook.

C4 (Core): Headline, accessibility, phone layout and export¶

Part A: Message titles¶

1. Rewrite the Trend title as a message title: what should the reader notice about 2025 against 2024? Use the month count from C3 Part C. The title must be true for the data shown.
2. Rewrite the Returns by region title as a message title that names the region your orange bar is about.
3. Double-click the dashboard title and write your headline: your big idea in one sentence, with a point of view and what is at stake. Look back at your big idea from the Session 14 lab.

Test the titles after a click. Click each region in turn. Is the trend title still true for every region? If not, rewrite it so it is true for every selection, or say in the title that it describes all regions. The region chart is not filtered, so its title does not change.

Part B: Accessibility check¶

Go through your dashboard with this list. Write yes, no or fixed next to each line in your log.

• The headline and the tiles can be read without looking at any chart.
• No more than five views (the four tiles count as one).
• Each colour means one thing everywhere: grey = context, blue = 2025, orange = needs attention.
• Nothing is shown only by colour: the years are written at the line ends, and the orange bar has its value and a title that names it.
• No red–green pair is the only difference between two things.
• Text is large enough to read when the page is shown at half size.
• The trend's axis starts at zero and the bars start at zero.
• The scope note gives period, units, source, how to use the click, and the main limit.

See it without colour. Export the dashboard (Part D) and view the PDF in greyscale (print preview in black and white, or your phone's greyscale setting). Can you still find the region your title is about? How?

Part C: Phone layout¶

1. At the top of the Dashboard pane, click Device Preview.
2. Under Device type, choose Phone, then click Add Phone Layout.
3. In the phone layout, keep the headline, the tiles and the region chart. Remove or move the other objects so nothing is too small to read.
4. Click Default at the top of the Dashboard pane to go back to the desktop layout.

Write one sentence in your log: which view did you leave out of the phone layout, and why?

Part D: Export and save¶

1. Click any selected bar again so that no region is selected.
2. With the Decision dashboard tab open, choose File → Print to PDF.... Under Range, choose Active Sheet, and click OK. Save as T3_lab_YourName.pdf.
3. Open the PDF and check that the headline, tiles, titles and scope note are all readable.
4. Save your workbook (File → Save).

---

Stretch tasks¶

S1 (Stretch): A dashboard for the marketing case¶

In the Session 14 lab, the marketing case asked: which channels deserve more budget next year, and which need a review? Build a one-page Tableau dashboard that answers it. Re-read the Session 12 lecture first if you need to.

Step 1: Download the file¶

The marketing-v1 files are on this course's public GitHub page, the same place the Python notebooks download them from.

1. Open the marketing-v1 folder on GitHub.
2. Click README.md and read the field list and rules. Note the unit of the money columns.
3. Go back and click marketing_daily.csv. Click the Download raw file button (a downward arrow, at the top right of the file view).
4. Save it in your course data folder. Keep the name marketing_daily.csv.

Step 2: Connect and check¶

1. In your T3 workbook, choose Data → New Data Source.
2. Under To a File, click Text file and open marketing_daily.csv.
3. Look at the field types in the preview. date must show a calendar icon. If it shows Abc, click the icon and choose Date.
4. Go to a new sheet. In the Data pane, check that the marketing data source is selected at the top. The fields may appear with tidied names, for example Attributed Revenue for attributed_revenue. Use the names you see in the formulas below.
5. Check the grain. README says one row per day and campaign. Drag the count field (marketing_daily.csv (Count)) onto Text. It must equal 365 days × 7 campaigns. Then drag campaign_id onto Text, right-click its pill and choose Measure → Count (Distinct): it must give 7. Do the same with date: it must give 365.

Step 3: Create the measures¶

README says rates are ratios of sums and must never be averaged. Create:

Field	Formula	Format
ROAS	SUM([attributed_revenue]) / SUM([spend])	Number, 2 decimals
CPA	SUM([spend]) / SUM([conversions])	Number, 2 decimals
Review Case	[spend] = 0 AND [conversions] > 0	(a True/False field)

In plain words: ROAS is attributed revenue per unit of spend; CPA is spend per conversion. A row with no spend but some conversions is a review case (README): it cannot have a ROAS of its own. Put Review Case on Rows and the count field on Text, and write the number of review rows in your log.

Step 4: Build the sheets¶

Sheet	How to build it
Four tiles: M Tile Spend, M Tile Revenue, M Tile ROAS, M Tile CPA	As in C1 Part A, with SUM(spend), SUM(attributed_revenue), ROAS and CPA. Third line: the unit (currency units, revenue per unit of spend, spend per conversion)
ROAS by channel	channel on Rows, ROAS on Columns, sorted descending, ROAS on Label. Add SUM(spend) to Tooltip, then right-click it and choose Quick Table Calculation → Percent of Total to show each channel's share of spend
CPA by channel	The same, with CPA
M Trend	Continuous MONTH(date) on Columns (the second Month, green). Drag spend to Rows, then drag attributed_revenue onto the same axis so Tableau shows Measure Values with Measure Names on Color. Label the line ends

Use the same colour rule as the Core dashboard: grey for context and one highlight colour for the channel your message is about, in both channel charts.

Step 5: Lay out and add the actions¶

1. Create a dashboard named Marketing dashboard, Fixed size, laid out like the Core dashboard: headline, tiles, trend, then the two channel charts side by side in a horizontal container, then a scope note. The scope note gives the period (2025), the unit (currency units), the source (synthetic marketing-v1), what attributed revenue does not show, and the review cases.
2. Add a filter action: source ROAS by channel, run on Select, targets the four tiles and M Trend, clearing shows all values.
3. Add a highlight action: Dashboard → Actions... → Add Action → Highlight.... Source sheets: ROAS by channel and CPA by channel. Target sheets: the same two. Run on Select. Under Target Highlighting, choose Selected Fields and tick channel.

Click a channel in the ROAS chart. The tiles and trend filter to that channel, and the same channel lights up in the CPA chart.

Step 6: Reconcile and read¶

• Click each channel and write its Spend tile. Do the five add up to the Spend tile with no channel selected?
• Divide total attributed revenue by total spend with a calculator. Does it equal the ROAS tile with no selection?
• Take the simple average of the five channel ROAS values. Is it the same as the overall ROAS? Which is right, and why?
• Find the channel with the highest ROAS. What is its share of spend? Would you move most of the budget there? Use the CPA chart and the scope note in your answer.

Write a message headline for the marketing dashboard. It must not claim that a channel caused revenue. Export the dashboard as a PDF too: T3_lab_YourName_marketing.pdf.

---

Challenge tasks¶

X1 (Challenge): Critique a poor dashboard¶

A colleague built this dashboard for a regional review and asks for your feedback. It has at least eight problems from this course. Read the description carefully.

Title: Dashboard

#	Component	What it shows and how
1	Six KPI tiles across the top	Sales, Profit, Quantity, SUM(Discount), Orders, and Avg returns share. Each tile is a red or green box with one number and no label for the period. The Orders tile uses the Orders (Count) field
2	Avg returns share tile	Calculated with WINDOW_AVG over the four region shares
3	Sales by region bar chart	Four bars, each a different bright colour. The value axis starts well above zero (Include zero is unticked), so the smallest region's bar is a stub
4	Sales trend line chart	Only December 2024 to February 2025, filtered by hand. Orange line. Title: Sales are falling fast
5	Region filter control	On the right edge. Apply to Worksheets is set to Only This Worksheet, on the trend sheet
6	Sales by sub-category pie chart	One slice for each sub-category, each in its own colour, with a legend
7	Profit by state map	Coloured from red (loss) to green (profit)
8	Ship days table	Every order with its ship mode and ship days, scrolling
9	A floating colour legend	Placed over the top right of the trend chart
—	Scope note	None. No units anywhere on the page
—	Size	Automatic

For each problem, write one row in your log:

Problem	Which rule or check it breaks	Your fix

Then answer:

1. What would the leadership team wrongly believe from the trend panel and its title? What would a fair comparison be? Hint: the Session 7 and T2 growth labs.
2. Which two tiles are wrong as measures, not only as design? What should each one use instead?
3. A reader picks West in the Region filter. Which views change and which do not? What would they wrongly conclude?
4. Sketch your redesign: one headline, at most five views, and where each sits. Which components did you remove completely?

X2 (Challenge): Prove two of the problems in Tableau¶

Use your own Decision dashboard workbook to show that two of the problems change the numbers, not only the look.

Problem A: the average of shares.

1. Duplicate Returns by region and name the copy Average of shares.

2. Create a calculated field named Avg of Region Shares:

   WINDOW_AVG([Returned Share 2025])
   

   In plain words: the simple average of the share bars in the view, each region counting the same.

3. Drag Avg of Region Shares onto Label. Right-click it and choose Compute Using → Region.

4. Compare the label with the Returned Share 2025 tile with no region selected. Which is larger? Use the Orders 2025 values from C3 Part B to explain why, in two sentences: which regions have the most orders, and what weight does the simple average give them?

Problem B: a filter that applies to one sheet.

1. Go to the Trend sheet. Drag Region to the Filters shelf, keep all regions, and click OK. Right-click the pill and choose Show Filter.
2. On the dashboard, click the Trend sheet, open its drop-down menu and choose Filters → Region to add the filter control to the page.
3. Pick one region in the control. Which views change? Write down what the tiles say and what the trend says. Do they describe the same selection?
4. Open the filter control's drop-down menu and choose Apply to Worksheets → Selected Worksheets.... Tick the tiles and Trend. Pick the region again. Now which views change?
5. Remove the filter control and the Region pill from the Filters shelf, so your submitted dashboard uses only the filter action.

Optional, Problem C: the truncated axis. On a duplicate of the trend or region chart, right-click the value axis, choose Edit Axis... and untick Include zero. Write one sentence on how the picture changes. Then delete the duplicate.

---

Evidence checklist¶

Before you submit, check that you have:

• the four tile sheets, Trend and Returns by region, each with a title
• the Decision dashboard with a headline, fixed size, scope note and the Filter by region action
• your region reconciliation table and the five checks from C3 Part B
• your trend reconciliation and month count from C3 Part C
• your accessibility check and your sentence on the phone layout
• the dashboard PDF
• Stretch and Challenge sheets and notes, if you did them

---

Submit on Moodle¶

1. Save your workbook as T3_lab_YourName.twbx (File → Save).
2. Export the Decision dashboard as a PDF (C4 Part D): T3_lab_YourName.pdf. If you did the Stretch task, export the Marketing dashboard too.
3. Open the T3 lab assignment on the course Moodle page.
4. Upload the .twbx file, the dashboard PDF (or PDFs), and your written answers (from your evidence log).

---

Next¶

Keep this workbook. In T4 you learn what Tableau Public makes public, what is safe to publish, and how to turn a dashboard into a portfolio piece from your capstone analysis.