Skip to the content.

Day 3 — Visualizing, Publishing & Automating Insights

Theme: From data to decisions — advanced reporting and storytelling

Prerequisite: Day 2. You need a working Power BI model to present — use Data Modeling.pbix or your Day 2 homework file.


The shape of the day

Days 1 and 2 were about making the numbers correct. Day 3 is about making them matter.

The morning covers visual design and narrative — how to choose charts that don’t mislead, and how to turn a page of figures into an argument someone acts on. The afternoon covers getting it in front of people and keeping it fresh without you, then a fast, gamified capstone — six timed rounds against genuinely messy data.

Power BI workflow: Sources → 1 Power Query, 2 Modeling, 3 DAX, 4 Visuals, 5 Share → Managers.

Today you finish the last two stages — Visuals and Share — and put the whole pipeline in front of decision-makers.


Morning Session — visualisation & storytelling

# Module Time You will learn
1 Visualization Best Practices 75 min Choosing the right chart, the axis-at-zero rule, conditional formatting, custom and AI visuals, slicers, drill-through, bookmarks, layout, accessibility
2 Storytelling with Data 45 min The three-act structure, titles that state findings, giving numbers a comparison, writing the recommendation, presenting

Afternoon Session — publishing & capstone

# Module Time You will learn Files
3 Publishing, Sharing & Automating 60 min Power BI Service, workspaces vs apps, licensing, scheduled refresh, gateways, row-level security, export, Analyze in Excel, Teams & SharePoint uses your Day 2 model
4 Capstone — The 45-Minute Dashboard Dash 45 min Six timed rounds — clean, model, measure, build and pitch a one-page executive dashboard against the clock, revising the whole Power BI thread novatech_capstone.xlsx

Capstone data: novatech_capstone.xlsx — NovaTech Retail’s raw export, download this before the afternoon session Optional Boss Level: raw_tables.xlsx (23 sheets, take-home) with reference build Project Clean Model.pbix Worked solutions (open after your own attempt): Dashboard Dash · Boss Level

Homework

Extend, Automate, Publish — about 45 minutes. Add a measure and a slicer-driven visual to your capstone, make it robust with a refresh indicator and RLS, then publish it.

No new files — you work in your own capstone build.


Expected outcome

You can design and present a fully functional Power BI dashboard integrating Excel data, demonstrating automation and insight delivery — and explain what should be done about what it shows.


The capstone is the real assessment

novatech_capstone.xlsx is nine sheets of NovaTech Retail’s raw export — the same business you’ve worked with since Day 1, now in the state real data actually arrives in: a duplicated backup sheet, two order years with mismatched columns, junk rows that quietly break a relationship, dates carrying a time component, and product names that don’t join until they’re trimmed. Three of the nine sheets don’t belong in your model at all, and knowing which is Round 0.

Every one of those maps to something taught on Days 2 and 3. Nothing in it is new — the challenge is doing it all together, at speed, and finishing with an argument rather than a report. There is a real, specific finding buried in the numbers, and the exercise only ends when you’ve said it out loud.

Deciding what to leave out is part of the exercise — and the built-in story rewards you for looking, not just building.

The full two-hour, 23-sheet version survives as an optional Boss Level (raw_tables.xlsx) at the end of the lesson, for anyone who wants to take it home.


Where the points are

Round Points
🕵️ 0 · Trap Hunt 10
🧹 1 · Clean 25
⭐ 2 · Model 15
🧮 3 · Measure 20
📊 4 · Build 25
🎤 5 · Pitch 5

Cleaning and modelling together are 50 of the 100 points. A beautiful dashboard on a broken model scores badly here — as it should, because in real life people believe it.


Previous: Day 2 · Course home: README


© 2026 Global Academy. Prepared for the ZIMASCO (Kwekwe) workshop. Facilitated by Tapiwa Zireva. Licensed to participants for personal learning — not for redistribution, resale, or reuse in other training without permission.