Five Ways to Present the Same Dataset (and When to Use Each)

Most teams have one dataset and one default way of presenting it — usually whichever format the tool they already use happens to produce. That's a habit, not a decision. The same underlying numbers genuinely need different treatment depending on who's looking at them and what they're about to do next.

To make this concrete, here's one dataset — Northwind's Q3 revenue by region, the same table this blog keeps coming back to — taken through five different presentations, with the actual reasoning for when each one is the right call.

select
  region,
  date_trunc('month', order_date) as month,
  sum(unit_price * quantity * (1 - discount)) as revenue
from orders
join order_details using (order_id)
join customers using (customer_id)
where order_date between '2025-07-01' and '2025-09-30'
group by region, month
order by month

Every version below reads from this same query. Nothing gets recomputed or copy-pasted between them.

1. The exploration canvas

An open, filterable view — every region, every month, drill-down available, nothing hidden. This is the right format when the audience is you, or anyone else actively trying to find something they don't already know is there.

The failure mode here isn't the format — it's sending this to someone who wanted an answer, not a search tool.

2. The single annotated report

One chart, one takeaway, one sentence of "why." No filters, no exploration — a fixed point of view for someone who needs the conclusion in ten seconds. This is what the KPI dashboard should become before it reaches a stakeholder who isn't going to dig.

3. The presentation

The same data, broken into a sequence — one region or one theme per slide, building toward a conclusion instead of stating it all at once. Right for a meeting where you're walking someone through reasoning, not just handing them a result.

4. The live embed

The same chart, dropped directly into a wiki page, a status doc, or this very blog post — staying current without anyone re-exporting anything. Right for anything that needs to still be accurate next month without manual upkeep.

5. The static export

A PDF or PNG, frozen at a specific point in time. This is the one format that's correct precisely because it doesn't update — an investor update for Q3 2025 should still say what Q3 2025 looked like, even after Q4 numbers come in. Right for anything that's a record of a moment, not a reference to an ongoing state.

The actual point

None of these five is the "correct" way to present revenue data — they're answers to five different questions (explore it, summarize it, walk through it, keep it current somewhere, or freeze it in time). The reason this is worth pointing out isn't that any one format is underused; it's that most tools force a choice between them upfront, so teams end up defaulting to whichever one the tool made default, regardless of which job actually needs doing.

Building this from one query rather than five separate exports is what makes switching between them cheap enough to actually do — and why we built Infigured around a single .infig document supporting all five as views of the same underlying object, rather than five separate files that quietly drift apart from each other over time.

---

Related: How to Present KPIs So People Actually Read Them