T4 lab: A portfolio piece from your capstone¶

Turn the strongest finding of your capstone into a portfolio piece: one Tableau dashboard and a three-point story that a stranger can follow in a minute. Save it on your own computer with a clear version history, check that everyone can read it, and, only if you choose to and only with safe data, publish it to Tableau Public.

This lab builds on your capstone from the Session 15 lab and your briefing from the Session 16 lab. You already have the decision, the evidence register, the checks and the recommendation. Today you show one part of that argument in Tableau. If you get stuck on a finding, look back at how you computed it in Python.

---

The business decision¶

Your capstone was written for the chief operating officer of the teaching frame. A portfolio piece has a different reader:

Hiring manager: "Show me one piece of analysis you are proud of. I have a minute. Tell me what question you answered, what the data showed, what it could not show, and what you recommended."

The decision is which one finding from your capstone best shows your work, and how to tell it so that a reader who has never seen this course understands it, trusts it, and knows its limits.

---

How this lab works¶

Part	What you do	Tasks
Setup	Choose your finding and its data, start a versioned workbook	Setup
Core	Plan the piece in four sentences; build one dashboard; build a three-point story; save and export	C1–C4
Stretch	Keep a version log; check the piece for accessibility	S1–S2
Challenge	Decide what is safe to publish; optionally publish to Tableau Public after the privacy checklist	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. Publishing is never required.

---

What you will be able to do¶

1. Choose one capstone finding and write it as question, evidence, limitation, action.
2. Build a one-message dashboard on your capstone data, with each dataset on its own data source.
3. Build a three-point Tableau story with a headline and a caption on each point.
4. Save versions of a .twbx with dated names and a change log.
5. Check a portfolio piece for accessibility.
6. Decide whether a workbook is safe to publish, using the privacy checklist.

Evidence you keep: your four-sentence plan, your saved workbook, the PDF of your dashboard and story, and any Stretch and Challenge notes.

Before you start, work through the T4 lecture. Have your capstone outline, evidence register and Session 16 brief to hand, and your T3 workbook for reference.

---

Setup: choose your finding and start the workbook¶

Step 1: Choose one finding¶

1. Open your capstone evidence register from the Session 15 lab (G2–G3).
2. Pick the one finding that carries your recommendation. It should come from one dataset.
3. Write in your log: the finding, its dataset, its metric, grain and period, and the check you ran in Python. You use the check again in C2.

Step 2: Download the data¶

Download only the file or files your finding needs.

Dataset	Where to get it	File to use
Superstore	Tableau's sample-data page, as in the Tableau overview	sample_-_superstore.xls (you already have it)
marketing-v1	marketing-v1 folder	marketing_daily.csv
finance-v1	finance-v1 folder	finance_actuals.csv, finance_budget.csv
operations-v1	operations-v1 folder	shipments.csv
hr-v1	hr-v1 folder	employees.csv, department_history.csv, recruitment.csv

For a synthetic dataset:

1. Open its folder on GitHub and read README.md first: the grain, the units and the rules for each measure.
2. Click the CSV file, then click the Download raw file button at the top right of the file view, as in the T3 lab, S1 Step 1.
3. Save it in your course data folder without renaming it.

Step 3: Start a versioned workbook¶

1. Open Tableau Desktop Public Edition.
2. Connect to your file: Microsoft Excel for Superstore (then follow Set up the Superstore data source), or Text file for a synthetic CSV.
3. Check the field types in the preview: every date must show a calendar icon.

4. Choose File → Save As, pick Tableau Packaged Workbook (*.twbx), and name the file with today's date and a version number:

   T4_portfolio_YourName_YYYY-MM-DD_v01.twbx
   

   For example T4_portfolio_AnitaShrestha_2026-11-20_v01.twbx.

---

Core tasks¶

C1 (Core): Plan the piece in four sentences¶

Before you open a sheet, write four sentences in your log. Use the T4 lecture, Part 5 and your Session 16 brief.

Part	Your sentence
Question	What decision, for whom?
Evidence	What the data shows, with a comparison, period and unit (no number yet: you fill it in from Tableau in C2)
Limitation	What the data cannot show; that the data is synthetic or a fictitious sample
Action	What should happen, who owns it, what to monitor

Then sketch the dashboard on paper, as in the T3 lab, C2: headline, tiles, the main view, at most one supporting view, and the scope note.

C2 (Core): Build the dashboard¶

Build the sheets first, then the dashboard. Reuse the methods from T1 to T3.

1. Create the measures your finding needs as calculated fields. Rates are ratios of sums, never averages of rates (T2 lecture, Part 2). Counts of orders, shipments or employees use COUNTD on the right ID.
2. Build two or three KPI tiles with a period label and a comparison (T3 lab, C1 Part A).
3. Build the main view that shows your evidence, for example a monthly trend or a ranked bar chart. Use grey for context and one highlight colour for the thing your message is about.
4. Reconcile. Compare each tile with the value in your Python evidence register. They must agree, apart from rounding. If they do not, find out why before you go on, and write the reason in your log.
5. Choose Dashboard → New Dashboard, name it Portfolio dashboard, and set Size → Fixed size first.
6. Lay out the page in reading order: tiles at the top, the main view in the middle, the scope note at the bottom.
7. Write a so-what headline (Dashboard → Show Title, then double-click the title) and message titles on each view. Each must stay true for every selection a reader can make.
8. Add a Text object at the bottom with the scope note: dataset and version (for example operations-v1, synthetic data for a fictional retailer), period, units, definitions of your measures, and the main limitation.

Save your workbook (File → Save).

C3 (Core): Build a three-point story¶

1. Choose Story → New Story. Name the tab Portfolio story.
2. Double-click the story title and write your headline: the same big idea as the dashboard.
3. Build three story points, each with a one-sentence caption:

Point	What it shows	Caption
1. Question and evidence	Your Portfolio dashboard	The question, and the finding with its number in context
2. A closer look	The main view, or a second view that tests the finding (for example the same measure by month, or one group against the rest)	What the closer look confirms, or where the finding is weaker
3. Limitation and action	The view that best supports the recommendation	What the data cannot show, then the action, owner and monitoring measure

1. To add a point, click Blank under New story point in the Story pane, then drag a sheet or dashboard onto the story. Use Duplicate to keep the same view with a different selection; click Update to keep a change in the current point.
2. Click through the three points in order. Read only the captions: do they tell the whole argument?

Save your workbook.

C4 (Core): Save a version and export¶

1. Choose File → Save As and save a new version: T4_portfolio_YourName_YYYY-MM-DD_v02.twbx. Keep v01.
2. Open the Portfolio dashboard tab, clear any selection, and choose File → Print to PDF.... Under Range, choose Active Sheet. Save as T4_portfolio_YourName_dashboard.pdf.
3. Open the Portfolio story tab and do the same: T4_portfolio_YourName_story.pdf.
4. Open both PDFs. Check that the headline, titles, captions and scope note are readable, and that all three story points appear. If the story PDF shows only one point, click each point in turn and export it on its own.

---

Stretch tasks¶

S1 (Stretch): A version log¶

Keep a change log in your evidence log, as in the T4 lecture, Part 3:

Version	Date	What changed	Why	Checked
v01
v02

1. Fill in v01 and v02 from today's work.
2. Ask a classmate to read your PDFs for one minute, then tell you the question, the finding and the action in their own words.
3. Make the change their answer shows you need (often the headline or one caption). Save it as v03 with File → Save As, and add the row to the log. The Checked column must say what you verified after the change.

S2 (Stretch): An accessibility check¶

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

• The headline and captions alone tell the argument.
• Nothing is shown only by colour: the highlighted mark also has a label, a value or 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; grey text is dark enough to read.
• Axes start at zero for bars.
• Every number has its unit, period and comparison.
• Tooltips are written in plain words, not raw field names such as AGG(ROAS). (Click Tooltip on the Marks card to edit them.)
• The scope note says the data is synthetic or a fictitious sample, and gives its source.

View the dashboard PDF in greyscale. Can you still find the mark your headline is about?

---

Challenge tasks¶

X1 (Challenge): Safe to publish?¶

For each case, write publish, do not publish or publish only after a change, and one sentence why. Use the T4 lecture, Part 2.

1. Your T3 marketing dashboard on marketing-v1, with a scope note that says synthetic data.
2. Your T3 decision dashboard on Superstore.
3. A dashboard of your employer's monthly sales by branch, with branch names replaced by letters.
4. A dashboard on hr-v1 whose scope note calls it "Our company's staff turnover".
5. A workbook on operations-v1 that also contains, on a hidden sheet, a second data source with Superstore.
6. A public dataset you found on a website with no licence, about house prices in your city.
7. A dashboard of course survey answers from your classmates, with names removed but with age and home town.

X2 (Challenge, optional): Publish to Tableau Public¶

Only do this if you want a public portfolio link. It is not needed to pass the lab, and nothing you submit on Moodle depends on it.

Step 1: Complete the privacy checklist¶

Synthetic course data may be published only once the course confirms its reuse terms, so ask your facilitator first. Write each line in your log with yes or no. If any line is no, stop: do not publish.

1. Every data source in the workbook is synthetic course data (marketing-v1, finance-v1, operations-v1 or hr-v1). Check the Data pane: every data source listed is published, even if no sheet uses it.
2. No data source holds workplace, customer or employee data, or anything copied from them.
3. Superstore is not in the workbook, not even on an unused or hidden sheet.
4. The scope note says the data is synthetic, describes a fictional organisation, and gives units (currency units) and period.
5. Nothing names a real person, a real company or your employer.
6. I have decided whether to allow downloads, and I know the views stay public either way.
7. I know that hiding the viz does not make it private, and that a downloaded copy cannot be recalled.

Step 2: Save a local copy first¶

Save your workbook locally with File → Save As (a new version number). Publishing does not replace your local file, and your Moodle submission is always the local .twbx.

Step 3: Publish¶

1. Choose File → Save to Tableau Public As….
2. Sign in with your Tableau Public account.
3. Enter a clear, descriptive title, for example "Peak-season delivery reliability (synthetic data)", and click Save.
4. Tableau creates an extract of your data and opens the viz on the Tableau Public website.

Step 4: Check what a stranger sees¶

1. On the viz page, open the viz settings. Decide whether to allow others to download the workbook and its data, and set the option. (Tableau's FAQ describes this option in the viz settings; the exact label may differ.)
2. Click Edit Details and write a short description: the question, that the data is synthetic and generated by a course for a fictional organisation, and the link to the dataset's folder on GitHub.
3. Copy the viz's URL. Open a private browser window (not signed in) and paste it. This is what a stranger sees. Check the download button and every tooltip.
4. Write the URL in your log.

---

Evidence checklist¶

Before you submit, check that you have:

• your four-sentence plan (question, evidence, limitation, action) and your paper sketch
• your reconciliation of the tiles with your Python evidence register
• the Portfolio dashboard with headline, message titles and scope note
• the Portfolio story with three points and a caption on each
• at least two dated versions of your workbook
• the dashboard and story PDFs
• Stretch and Challenge notes, if you did them (version log, accessibility check, X1 answers, the privacy checklist and URL if you published)

---

Submit on Moodle¶

1. Save your final version locally (File → Save As): for example T4_portfolio_YourName_YYYY-MM-DD_v03_submitted.twbx. Do not change this file after you submit.
2. Export the dashboard and the story as PDFs (C4).
3. Open the T4 lab assignment on the course Moodle page.
4. Upload the .twbx file, the PDF (or PDFs) of your portfolio piece, and your written answers (from your evidence log). If you published in X2, add the URL.

---

Next¶

This is the last Tableau session. Keep your portfolio workbook and its version log with your capstone notebook and brief from Session 15 and Session 16: together they show the whole path from a business question to a checked recommendation.