Presenting Your Analytical Work

BUS220 - Business Intelligence and Analytics | Week 7

Oleh Omelchenko

2026-04-22

Week 7 is a project week

“The original plan for this week was Tableau polish and performance — accessibility, tooltips, optimization. With Assignments 4 and 5 and the project all running at once, adding new tool features now would compete with work you already have to finish. So we’re doing something different. This week we review the whole course through the lens of presenting your project, and fill the one gap we haven’t addressed: how to actually present analytical work.”

What this week covers

No new Tableau mechanics this week. Instead:

• A framework for the whole arc from business question to presented story
• Tactics for the group project presentation
• A self-audit and peer review in practice

You’ve spent six weeks learning to build dashboards. Nobody has shown you how to present them yet. Let’s fix that.

“Presenting analytical work is the single biggest gap in this course. It’s also one of the most valuable skills for analytics careers — in a job, you spend more time presenting findings than building charts. The group project presentation next week is the reason to learn this now, but the skill carries forward.”

The project, briefly

Each group gets one city and two quarterly Airbnb scrapes (~80 columns per listing: pricing, host profiles, reviews, availability, location, licensing). Each member picks one analytical area and builds one dashboard page for it.

Pricing & Value

What drives pricing in this city’s rental market?

Host Profiles & Market Structure

Who runs this market — casual hosts or commercial operators?

Guest Experience & Reviews

What does the review data reveal about satisfaction and seasonal demand?

Geography & Regulation

How is the market distributed, and what does regulation look like?

You’ll hand in a Tableau workbook (KPI overview + one page per area), a written analysis, and a live presentation where each member defends their own area.

“Quick refresher in case anyone hasn’t opened the project brief yet. The data comes from Inside Airbnb — an independent project that scrapes public Airbnb listings. Each group is assigned a different city, and each group gets two quarterly snapshots so you can compare across time.”

“The four areas are intentionally split so each group member has their own analytical territory. You collaborate on data prep, you share a KPI overview page, and then each member owns their area’s page and defends it individually.”

“When I say ‘your area’ in the rest of the lecture, I mean whichever of these four you’re working on. When I say ‘your project,’ I mean the group’s collective work. Keep those separate in your head.”

“Everything we’ll talk about today applies to all four areas. The framework doesn’t care whether you’re on Pricing or Geography.”

After a presentation, people remember

63%

of the stories

5%

of the statistics

From Made to Stick by Chip and Dan Heath. If you remember one thing from today, make it this.

“Heath and Heath studied what people remember after presentations. The numbers are striking. Sixty-three percent remember the stories they heard. Five percent remember the individual statistics. This doesn’t mean statistics don’t matter — without them, the stories are just opinions. But numbers alone don’t survive the walk back to the office. The story is what carries the numbers with it.”

“This is why this week exists. You’ve been building numbers. You need to learn how to wrap them in something that survives.”

Sizzle and steak

Tableau runs an annual competition — Iron Viz. Three finalists, same dataset, 20 minutes to build live.

Ryan Sleeper won in 2013 and later wrote Practical Tableau, the book we’ll borrow most of today’s structure from. In his own telling, both competitors outclassed him (one on analysis, one on design). He won on something else.

Steak

data, analysis, design

+

Sizzle

story, motivation, delivery

When it was time to explain the dashboard, the other two walked through technical choices. Sleeper told the story behind the analysis.

“The Iron Viz analogy comes from Ryan Sleeper — the 2013 winner, author of Practical Tableau. He had 20 minutes to build a dashboard on a dataset about movies. His concept — ‘Do Old Movies Get Better with Age?’ — was interesting but not technically more sophisticated than his competitors. He’s been open about saying one competitor was better at analysis, the other better at design.”

“He won because at the end, when each finalist got time to explain their dashboard, the other two walked through technical aspects and user experience. Sleeper told the story behind what motivated his approach. That’s what the judges remembered.”

“Keep this in mind. We’ll come back to it at the end of the lecture.”

A framework for the whole arc

Sleeper’s INSIGHT framework (from Practical Tableau, Part IV): seven steps that take you from a business question to work you’ve actually delivered.

I

Identify
question

N

Name
KPIs

S

Shape
data

I

Initial
concept

G

Gather
feedback

H

Hone
dashboard

T

Tell the
story

First three letters are strategy. Next three are execution. The last one — T — is where most of today goes.

“The framework comes from Part IV of Practical Tableau. It captures something most data analysts have to learn the hard way through years of iteration: if you don’t get the first three steps right, everything downstream is wasted effort.”

“We’re going to walk through all seven steps, applying each one to your group project. Strategy comes first — the I-N-S steps decide what the dashboard is for. Execution is the I-G-H steps — building, getting feedback, refining. Tell the story is the last step and the hardest one, because it doesn’t live in Tableau at all. It lives in how you communicate what you built.”

Step 1: Identify the question

I

N

S

I

G

H

T

“Strategy before building. The first step is deciding what the dashboard is for.”

One dashboard, one question

A dashboard that tries to answer everything usually ends up answering nothing clearly.

Sleeper’s running example is a dashboard titled:

Do Old Movies Get Better with Age?

Everything below the title is there to answer it.

“This is the title of Sleeper’s Iron Viz winner. Six words. A single question. Everything on the dashboard below — the average rating gauges, the trend line, the ratings broken down by age group — exists to answer it.”

“The question goes literally at the top. Not as a subtitle, not as fine print — as the title. Why? Because you won’t always be there to explain the dashboard. Your audience has to know what it’s for within the first three seconds. The question does that job.”

“The word ‘question’ is singular on purpose. Rabbit holes will emerge. Follow-up insights will appear. That’s fine. But by going in with one question, you’re almost guaranteed the dashboard serves at least that one purpose. Try to answer five questions at once and you’ll likely serve none of them.”

The movie dashboard

Question at the top. Under it: average rating for new vs. old movies, a trend line with a slider, demographic breakdowns at the bottom. Every piece is there to answer the title.

“Read the structure from top to bottom. The question is the headline. The two big numbers — 7.60 for new movies, 8.18 for old movies — give a first-pass answer. The trend line shows how rating changes with age, and a slider lets the user set their own threshold for ‘old.’ The bottom row breaks it down by the reviewer’s age and gender.”

“None of this is decoration. Every element connects to the question. If a chart doesn’t help answer ‘do old movies get better with age,’ it doesn’t belong on the dashboard. That’s the discipline.”

From generic to sharp

A coffee shop chain asks for a sales dashboard. A stakeholder says:

“How are sales doing?”

Too vague. “Doing” compared to what? Which locations? Which products? No dashboard can answer a question this broad.

“Which product categories are driving the weekday afternoon sales dip at our downtown locations?”

Specific, answerable, short enough for a title.

“The generic version could describe a thousand dashboards. The sharp version describes one. Notice what the sharp version commits to: specific segment (product categories), specific time window (weekday afternoons), specific locations (downtown). The dashboard below it has a job to do.”

“Generic questions come from stakeholders who don’t yet know what they need. Part of your job is turning their ‘how are we doing’ into something answerable.”

Your turn: refine your question

Write down your current business question for your area.

Then apply the title test:

• Does it fit in one line?
• Would someone who’s never seen your project understand what you’re asking?
• Is it specific enough that a different city’s data couldn’t answer it?

Any “no” means it needs another pass.

“Pause here for 60 seconds. Give students silence to write. Don’t narrate over the thinking time.”

“After the minute, ask: who wants to read theirs out loud? Pick one volunteer. Apply the three tests together in front of the class. Don’t criticize — ask questions. ‘Does it fit in one line?’ If no, ‘what part is load-bearing, what part is optional?’ If yes, ‘does the dataset actually answer this?’”

“If no volunteers, ask a specific group: ‘Pricing team — read me yours.’ Most groups will be willing once one does.”

“This is the first time in the lecture they apply the framework to themselves. Keep it moving — 2-3 minutes total, then keep teaching.”

Step 2: Name the KPIs

I

N

S

I

G

H

T

“Once you have the question, you need the metrics that answer it. That’s step two.”

Five or six, no more

Sleeper’s rule of thumb: five or six metrics at the top, no more. Past that and you haven’t decided what actually matters.

$3.2k

Weekday PM sales

−18%

vs. morning

42%

Category share: bakery

214

Tx per hour

Primary KPIs — always visible

SKU breakdown

customer segment

active promo

weather

staff on shift

avg basket

…

Secondary — tooltips & drill-downs

If everything is “key,” nothing is.

“The temptation is to stack every metric you can calculate. A host count here, a median price there, a superhost share, a response rate, a review score, a calendar availability, a scrape date range, an estimated revenue — and now you have twelve KPI cards at the top, none of which the audience can take in at a glance.”

“5-6 primary KPIs is the discipline. Everything else moves to a secondary view. Tooltips, drill-downs, a detail table, a separate page. They’re not gone — they’re just not competing for your audience’s first attention.”

“This is worth explicitly discussing with your group. Get alignment on the primary KPIs before you build.”

KPIs ladder up to the question

The movie dashboard asks: Do Old Movies Get Better with Age?

Three KPIs:

• Age — the question’s central variable
• Average rating — the measure of “better”
• Number of reviews — the sample size check, because old movies have fewer reviews

Three metrics, each tied straight back to the question. Nothing else on the dashboard needs KPI-level prominence.

“Notice what’s not on the list. Box office revenue. Genre. Director. Country of origin. Oscar wins. All interesting, all potentially related to movie quality — but none of them are the question. They’d add noise.”

“When you design your KPIs, the test is: does this metric directly help answer the stated question? If it’s a nice-to-have, it belongs in a secondary view.”

“Think about this for your own area. What are the three to five metrics that are truly primary? The rest go to detail views — they’re still valuable, just not competing for the KPI row.”

Back to the coffee shop

Question: “Which product categories are driving the weekday afternoon sales dip at our downtown locations?”

Primary KPIs

• Sales by category × hour of day
• Category share of total
• Weekday vs. weekend comparison
• Transactions per hour

Secondary metrics

• SKU-level breakdown
• Customer demographics
• Active promotions
• Weather conditions

Four primary KPIs answer the question. Everything else lives one click away.

“Four primary KPIs for a question about the afternoon sales dip. Why these four? Sales by category × hour shows where and when. Category share shows what’s proportionally dropping. Weekday vs. weekend tests whether it’s a Monday-Friday pattern. Transactions per hour separates ‘fewer customers’ from ‘smaller baskets’ — a key analytical fork.”

“The secondary metrics are useful, but don’t belong in the KPI row. If someone asks ‘is weather causing it?’ they click into a detail view. If someone asks ‘which SKU is worst?’ — drill-down. The primary row stays focused.”

Your turn: name your KPIs

For your area’s refined question, draft:

3-5 primary KPIs

Metrics that directly answer the question.

3-4 secondary metrics

Useful context for drill-downs or tooltips.

Compare with your teammates after the lecture. Align before building.

“90 seconds of silent writing. Don’t narrate.”

“The goal isn’t to finalize the list during the lecture — it’s to start the conversation. Groups will argue about their primary set because different members see different priorities. That argument is the work. Better now than after three evenings of building.”

“After the 90 seconds, don’t collect answers out loud — KPI lists are boring to hear from the podium. Instead, ask: ‘Was this easy or hard?’ Most will say hard. That’s the point — if the question is sharp, the KPIs follow. If the question is vague, the KPI list is endless.”

Step 3: Shape the data

I

N

S

I

G

H

T

“Once you know the question and the KPIs, the third step is getting the data into a form you can work with. You’ve been doing this all trimester — here’s how to think about it as part of a process.”

Shape the data you have

Two parts to this step:

1. Layout — vertical tables, one field per column, one entry per row (Tidy Data from Week 3).
2. Transformation — find the fields you need, even if they aren’t obvious in the raw data.

Most of the fun is in the second part.

“Layout is straightforward — you’ve done it for ten weeks. Every dataset we’ve touched needed reshaping before Tableau could work with it. Pivot tables to tidy format, unpivot survey data, stack two scrapes into one.”

“Transformation is the creative part. Sometimes the field you need isn’t in the data — but it’s hidden inside another field. That’s when you get creative.”

Finding age inside a movie title

Sleeper’s Iron Viz dataset didn’t have an age column. But the year of release was sitting right there in every title — a field hiding inside another.

Casablanca (1942)

→

extract year

1942

→

2026 − year

84

Shape the data to fit your question. When a field you need isn’t there, you can usually derive it.

“Sleeper’s dataset came from Tableau as-is. Title, genre, rating, year would have been nice — but year wasn’t a clean field, it was embedded in the title with parentheses. A calculated field with MID and FIND extracted the four-digit year. Another field subtracted it from the current year to get age.”

“This is exactly the kind of thinking you’re doing on the project. Price arrives as ‘$169.00’ — a string. Bathrooms come as ‘text’ — ‘1.5 shared baths.’ Amenities arrive as a JSON array. Each of these needs a small transformation before you can use it. Assignment 5 walks through the process on Edinburgh.”

“The creative analyst isn’t the one who complains about dirty data. It’s the one who looks at what’s in the data and figures out how to get what they need out of it.”

Step 4: Initial concept

I

N

S

I

G

H

T

“With the question, the KPIs, and the data in shape, you could open Tableau now. Don’t. There’s one more step first.”

Sketch first

The temptation is to open Tableau and start dragging fields around.

Sleeper’s case: sketch on paper first. Boxes, labels, stick-figure level. No artistic skill required.

Two reasons:

• It’s fast. Three layouts in ten minutes, easy.
• It’s cheap to throw away. A critique of a sketch barely stings. A critique of a polished dashboard you’ve poured two evenings into — that one stings.

“This is counterintuitive for students who are comfortable in Tableau. Why sketch when you could build?”

“Because your concept will change. It always does. The first person who sees it will have ideas. Your teammates will have ideas. You will have ideas tomorrow that you don’t have today. If you’ve built a polished dashboard, those ideas feel like losses — you’ve already spent the effort. If you’ve sketched, they feel like wins — the sketch was cheap and the next version will be better.”

“This is about protecting the quality of the feedback you get, not about saving time in Tableau.”

A real sketch

Sleeper’s planning sketch for an e-commerce dashboard. Boxes, labels, rough proportions. Current week vs. last year. Metric switcher at the bottom. Six KPIs at the top, four more below, a main trend chart at the bottom.

“Look at what this sketch communicates even at its ugly resolution. You can tell what’s at the top, what’s in the middle, what’s at the bottom. You can tell there are two metric groups — eCommerce KPIs and Merchandising KPIs. You can tell there’s a chart across the bottom that uses a filter control.”

“That’s all a sketch needs to do. It communicates layout intent, not polish. If you showed this to a stakeholder, they could react to the structure without getting distracted by specific chart types, color choices, or exact numbers. That’s exactly what you want at this stage.”

“Many of you will feel self-conscious about sketching. It’s OK. Nobody’s grading the sketch. The sketch is a working document — it exists to provoke the right conversation at the right time.”

Audience and purpose drive the concept

Two questions to answer before you sketch:

1. Who is the audience?
2. What is the purpose?

For your project:

• Audience — the presentation panel, your teammates, readers of the PDF
• Purpose — answer the analytical area’s business question, with specific recommendations

A sketch for a C-suite audience looks nothing like one for fellow analysts. Know who’s on the other end before you draw a box.

“Sleeper distinguishes three audience types: analysts (who can handle sophisticated chart types and dense views), mass audiences (who need simplified layouts and clear stories), and C-suite audiences (who want insights boiled to the top, with prescription in view).”

“For the project, your audience is mixed but leans toward the mass/analyst end — the panel has seen a lot of Tableau, but not your city’s data. Your teammates are in the same cohort as you. Readers of the PDF might not have the presentation context.”

“Design for the slowest reader in your audience — someone picking up the PDF cold. If the slowest reader can understand it, the panel can too.”

Step 5: Gather feedback

I

N

S

I

G

H

T

“Once you have a sketch, don’t go build. Get feedback first.”

Three reasons to get feedback early

Benefit	Why
Efficiency	Organized early feedback beats scattered late feedback
Reduced frustration	It’s easier to hear “change this” about a sketch than about a finished dashboard
Ownership	People who influenced the concept defend the final product

Build with your audience, not for them.

“The efficiency argument is obvious. You don’t want to build twice.”

“The frustration argument matters more than people admit. When someone critiques work you’ve invested two evenings in, it hurts even when the critique is right. When someone critiques a sketch, it barely registers. Sketches are cheap to change. Use that to your advantage.”

“The ownership argument is the quiet one. People who influenced the dashboard treat it as partially theirs. They defend it, they explain it to others, they push to make it succeed. People who were shown a finished artifact often treat it as the analyst’s problem. For the project, this matters within the group — if one person decides everything and shows finished pages to the others, the others won’t defend the work in the presentation. If everyone had input at the concept stage, everyone has skin in the game.”

For the project: review inside the group

Your group is the feedback loop. You’re already sharing the data prep and the workbook — teammates will spot things a solo builder can’t.

• Rotate sketches across the four members before anyone opens Tableau. One person sketches their page; the other three react.
• Rotate again after the first build. Fresh eyes catch what the author no longer sees.
• Before polishing, run the INSIGHT audit against each page with a teammate.

Collaboration inside the group is part of the project design. Workbooks, calculated fields, and the written analysis all stay inside the group.

“The group-internal loop is free and high-value. It takes 20 minutes per member to share a sketch and collect reactions. Do it before anyone builds. And do it again after the first build — different things become visible once the dashboard is real.”

“Your teammates are the best reviewers you have. They know the data, they share the analytical goals, and the project design treats their collaboration as legitimate. Use them.”

“Academic integrity boundary: the workbook, the calculated fields, and the written analysis stay inside your group. Cross-group workbook exchange would put you in a grey zone none of us want to navigate. There is no shortage of feedback available inside the group — that’s where it belongs.”

Step 6: Hone the dashboard

I

N

S

I

G

H

T

“Now you build. Most of your Tableau time lives here. A few rules to keep the work focused.”

Rules for the build phase

From PT Ch 82:

• Dimensions match distribution. Printed handout = paper size. Laptop presentation = laptop resolution. Mobile = phone size.
• 12 components maximum. KPI cards, charts, filters — all counted. If you need more, split to another page.
• Keep it simple. Chart selection, user experience, design. Obvious when said. Violated constantly.

If a view requires three sentences of explanation, it probably belongs in two views, not one.

“Twelve components sounds generous until you count. Four KPI cards plus three charts plus two filters is nine. Add a breakdown table and a legend and you’re at eleven. One more chart and you’re over budget. This is a useful constraint — when you’re at 12, any new view has to replace an existing one.”

“Keep It Simple is the rule nobody follows. Students often default to complex chart types because they’re impressive. Bubble charts, Sankey diagrams, radial layouts. If the same insight can be shown with a bar chart, show it with a bar chart. The audience is not grading your chart catalog.”

“For the project specifically: resist the urge to put everything on one page. Four members means four area pages plus a KPI overview page. Don’t try to cram all four areas into one mega-dashboard. That’s what navigation is for — you covered it in Week 6.”

When feedback conflicts with best practice

A stakeholder wants 15 pie charts. Best practice says no. What do you do?

Sleeper’s move: build both. The one they asked for, and the one you recommend. Let them compare. Let them decide.

You’re the final gatekeeper — but you’re also not the user. When you disagree, split the difference.

“This comes up on the project more than people expect. One teammate wants a complicated visualization because it’s interesting. Another prefers a simple bar chart. A member of the panel might have specific preferences you don’t share.”

“The ‘provide both’ strategy is diplomatic and honest. You’re not overriding them; you’re showing them the alternative. Often, the comparison alone convinces. When it doesn’t, at least you’ve made your recommendation visible — you’re not complicit in a decision you disagreed with.”

“In the written analysis, you can document your design choices. That’s where you explain why you went with what you went with.”

Step 7: Tell the story

I

N

S

I

G

H

T

“We’ve reached the step nobody is taught in Tableau courses. The step that separates the Iron Viz winner from the other finalists.”

Two ways to persuade

Robert McKee teaches screenwriting (his students include 60+ Academy Award winners). Harvard Business Review interviewed him specifically about business persuasion — why some pitches land and others don’t. He drew this distinction:

The first is conventional rhetoric — slides filled with bulleted facts and statistics. If you succeed in persuading them, you’ve done so on an intellectual basis. That’s not good enough, because people are not inspired to act by reason alone.

The second way is through story. Stories unite an idea with an emotion.

“McKee is a screenwriting lecturer — his former students include sixty-three Academy Award winners. He was interviewed for Harvard Business Review specifically about persuasion in business, and this is the distinction he drew.”

“It’s not that rhetoric is wrong. It’s that rhetoric is necessary but not sufficient. You can convince someone intellectually that your recommendation is correct and still fail to move them to act. Action requires engagement — and stories are how humans engage.”

“For data work, this is the claim that justifies this whole block. If you present your findings as a list of statistics, you’re in the 5% zone. If you present them as a story with structure and tension and resolution, you’re in the 63% zone. Same data. Different delivery. Different outcome.”

Three parts of a data story

Storytelling	Data work
Characters	Your KPIs and contextual metrics
Plot	The insights you uncovered
Narrative	How you communicate them

Your analytical work already has the first two. The third — narrative — is what the presentation brings.

“A story has three components. Characters — who the story is about. Plot — what happens. Narrative — the voice telling it.”

“Map this to your project. The characters are your metrics: price, listings, host tiers, review scores. The plot is what you found — the neighbourhoods that underprice themselves, the commercial operators dominating specific areas, the seasonal patterns. The narrative is how you walk the audience through it.”

“Most of your work has been on characters and plot. This lecture is almost entirely about narrative.”

A useful pattern: setup → conflict → resolution

Aristotle sketched a three-act shape for plays, and 2,400 years later it still works. Business presentations borrow it as a mental model — not a rule, just a way to think about flow.

Setup

Establish the situation. Introduce the audience’s stake in it.

Conflict

Show what the data reveals. Build tension between what is and what could be.

Resolution

Deliver the recommendation. Tie back to the setup.

Use it when it helps. Drop it when it doesn’t.

“Three acts. Setup, conflict, resolution. Same structure as Red Riding Hood, Star Wars, and every TED talk you’ve sat through.”

“For a data presentation, this maps naturally: the setup establishes what you were asked to look at. The conflict is what the data shows — often something unexpected or inconvenient. The resolution is what to do about it.”

“I’m offering this as a pattern, not a requirement. Some presentations don’t need a traditional arc — a status update or a weekly operational report is fine without one. But when you’re presenting analytical findings to someone who has to decide something, this shape usually helps.”

“The key thing to hold on to: your presentation needs a beginning, a middle, and an end. Many presentations we’ve all sat through are missing one of those. Often it’s the end — the ‘so what’ — that gets cut for time, leaving the audience with ‘interesting data, now what?’”

Narrative flow: two orders

Chronological

Walk through in the order you did it.

Problem → data → analysis → finding → recommendation.

Works when you need to establish credibility, or the audience cares about the process.

Lead with the ending

Start with the recommendation. Then back into the evidence.

Recommendation → key evidence → supporting detail.

Works when you already have the audience’s trust, or they care about “so what” more than process.

“Both orders work. The question is which fits your situation.”

“Chronological is safer when the audience is skeptical, or when they’re analysts themselves who care about methodology. You’re saying: ‘Here’s how we got here, step by step.’ It builds credibility through the journey.”

“Lead with the ending is stronger when the audience is busy and wants the headline up front. Executive audiences often want this. ‘Here’s our recommendation. Now let me show you why.’”

“For the group project presentation: mix both. Open with a hook about the finding — that’s leading with the ending. Then cover the data prep and methodology — that’s chronological. Then close with the specific recommendations — that’s the resolution, linked back to the opening hook.”

The power of repetition

The three shapes represent your three main points. Preview them up front, walk through each one, recap them at the end. The audience hears each point three times.

INTRO   ▲  ●  ■

▲

●

■

CONCLUSION   ▲  ●  ■

Bing — tell them what you’re going to tell them

Bang — tell them

Bongo — tell them what you just told them

Coffee shop opening: “We’ll cover what the sales dip looks like ▲, which categories are driving it ●, and what to do in the afternoon menu ■.”

To you it feels redundant. To the audience — hearing it for the first time — it feels helpful.

“Cole Knaflic attributes this to her junior high English teacher. The name is silly and that’s why it sticks.”

“The structure is three repetitions of the key points. First as preview — ‘here’s what you’ll hear.’ Then as content — the actual analysis. Then as recap — ‘here’s what you heard.’”

“Presenters skip this because it feels patronizing — of course you don’t need to say the same thing three times, you’re presenting to adults. But the audience isn’t in your head. They’re hearing it for the first time, often distracted, often multitasking. Repetition is how you ensure they leave with the message.”

“For the project: open with ‘Our analysis of Vienna’s Airbnb market revealed three findings…’ (preview). Present each finding with evidence (content). Close with ‘To recap — we found X, Y, Z. Our top recommendation is…’ (recap). Three passes. Nobody leaves the room wondering what the point was.”

Four tests for story clarity

Before you present, run your deck through these:

• Horizontal logic — read just the slide titles in order. Do they tell the story on their own?
• Vertical logic — on each slide, does everything on it reinforce the title?
• Reverse storyboarding — flip through the deck, write the main point of each slide. Does the list read like an outline?
• Fresh perspective — show the deck to someone with no context. What do they think the story is?

Wherever one of these fails, that’s where your narrative needs work.

“Horizontal logic is the harshest test. Open your deck, hide the content, read only the titles in order. If that reads as a coherent story, your narrative is embedded in the structure — it doesn’t depend on you being in the room. If the titles are ‘Data Overview,’ ‘Price Analysis,’ ‘Host Analysis,’ ‘Conclusion,’ you have section labels, not a story.”

“Vertical logic is about each slide standing on its own. If the title says ‘Commercial hosts dominate the central district’ and the slide shows a chart plus four unrelated text boxes, the vertical logic is broken. Everything on the slide should reinforce the title.”

“Reverse storyboarding is a quick diagnostic. Takes five minutes. The result should look like the outline you would have written if you’d planned the deck first.”

“Fresh perspective is the most humbling. Find someone not in your group, show them the deck, ask what they think you found. If their answer matches your intent, you’re done. If it doesn’t, you have work to do — and you know where to focus.”

Worked example (if time permits)

“Time check moment. If we’re on pace, we walk through Knaflic’s comprehensive Chapter 8 example — it shows the whole arc end-to-end, same chart used eight times, each with different emphasis. That’s the most vivid illustration of what we’ve covered.”

“If we’re running tight, skip ahead. Point students to the book chapter as optional reading. Nothing downstream depends on this section — the takeaways slide still works.”

The setup

Cole Knaflic has a full worked example in Storytelling with Data, Chapter 8.

The scenario: you work for a startup. The team needs to price a new consumer product, and you have seven years of competitor pricing to work with.

The Big Idea — the analysis in one sentence:

Based on analysis of pricing in the market over time, to be competitive, we recommend introducing our product at a retail price in the range \(ABC–\)XYZ.

“The Big Idea statement is what the analysis is trying to communicate. Audience: VP of Product. What: understand competitor pricing over time, recommend a range. How: show retail prices for five competitors over seven years.”

“Everything that follows turns that raw data into that single sentence, delivered so the VP can act on it.”

Let’s open Chapter 8 together

We’ll skim it together. As we go, watch for:

• The first chart. Notice what’s wrong with it — the colors, the axes, the title that doesn’t match the visual.
• The stripping process. Each iteration removes something. Track what’s lost and what starts to appear.
• The story build. Once the chart is clean, the same chart shows up six more times with different emphasis. Notice what changes each time.
• The final slide. How does it tie back to the Big Idea?

The data never changes. The delivery does.

“Open the book on the projector and skim Chapter 8 from the original cluttered bar chart (Figure 8.1) through the final recommendation slide (Figure 8.19) and the before-after (Figure 8.20). This should take 6-8 minutes of live skimming, pausing at a few key figures to comment.”

“Key figures to dwell on: 8.1 (original — cluttered, conclusion title doesn’t match the visual), 8.4 (line graph conversion — turning point), 8.6 (clutter eliminated — clean canvas), 8.11 (goal slide — the Bing), 8.13-8.17 (the story build, one slide at a time), 8.19 (recommendation — the Bongo), 8.20 (before/after, two panels side by side).”

“At each figure, ask the class a quick question: ‘What changed between this and the previous one? Why?’ Most figures show one deliberate change — removing color, changing chart type, hiding lines, adding the recommendation band. Students should be able to name the change before you do.”

“If the class has trouble with this, slow down. If they’re confident, keep moving. The goal is that they feel the pattern of small, deliberate changes building a story — not that they memorize the sequence.”

What to export to your project

Four things worth carrying back:

1. State the Big Idea first. One sentence that names the audience, the finding, and the recommendation. Write it before you design any slide.
2. Strip the chart to its essentials. Pull out anything that isn’t carrying information. The clean chart is your canvas.
3. Reveal only what the current slide needs. Same chart, different emphasis each time. Hide the lines that aren’t part of this moment.
4. Close by tying back to the Big Idea. The last slide answers the question the first slide raised.

One chart used eight times with different emphasis beats eight different charts every time.

“This is the export. Read through Chapter 8 tonight if you want the full version — but the four rules above are the core.”

“The third rule is the most counterintuitive and the most powerful. Students often think they need a different chart for each slide. Knaflic’s example uses the same line chart for the entire story, just changing which lines are highlighted and which are greyed out. That’s easier to build and easier to follow.”

“For your project presentation: pick one core chart per area. Show it multiple times, each time highlighting a different aspect of the finding. The audience’s brain stays oriented because the chart is familiar — but the story advances because the focus shifts.”

The Iron Viz winner’s edge

Sleeper’s two competitors spent their explanation time on the technical side — how the calculation worked, why they picked the chart type, where they placed the filter.

Sleeper told the story of why he was curious about old movies in the first place.

For your defense: the panel already has the PDF. They know what you built. Tell them why you built it, and what it means.

“This is the payoff from the opening story. Sleeper didn’t walk through calculations. He told the story — why he chose the question about old movies, what he expected to find, what the data actually showed, what it means.”

“For your defense: the panel has read the PDF. They’ve seen the workbook. They know what’s in it. What they don’t know is what motivated your analytical choices, what surprised you, what you believe the data is saying beyond the charts.”

“That’s what your five minutes should be for. Not a click-through tour of your pages. A story about what the data showed and what you think should happen because of it.”

Deadlines

• Assignment 5 — due end of this week
• Group project workbook + PDF + data prep scripts — due day before Week 8 presentations
• Individual defense — mandatory at the presentation

See group-project.qmd for the detailed rubric.

“Quick deadline reminder. Assignment 5 is this week’s priority alongside the project. The project deliverables are due the day before presentations — don’t cut that close. Every group member defends their own area individually — not having someone present for their own area is handled per KSE academic integrity policy.”

What to take away

• A single question at the top. If it needs qualification, it’s not the question yet.
• Five or six KPIs. If everything is key, nothing is.
• Sketch before you build. Cheap to throw away.
• Feedback before polish. Protect the quality of the feedback you get.
• Stories stick. Statistics don’t — 63% vs. 5%.
• Tell them why you built it, and what it means. Skip the technical tour.

“Six takeaways. Print them, tape them to your monitor, apply them to the project this week.”

“The first four are about building better work. The last two are about presenting it. Most students spend 90% of their time on building and 10% on presenting. For this project, spend 30% on presenting. That’s the ratio that matches how you’ll be evaluated — and how your work will actually be remembered.”

“Good luck with Practice 1. See you tomorrow.”