Identify the reader's repeat task
An operations team may need to filter recent records, compare periods and return to the same view every week. That is a dashboard task. A public audience may need context, a sequence of evidence and an explanation of what a pattern can and cannot mean. That is a story task. If both audiences matter, do not force one interface to carry both jobs without an explicit transition.
Start with three questions a real reader will ask. For each, record the data source, unit, time period, missing values and decision the answer supports. This prevents a visually striking chart from implying certainty that the underlying records do not provide.
| Question | Dashboard | Data story |
|---|---|---|
| Who controls the path? | The reader filters and revisits | The author proposes a sequence |
| How often does it change? | Usually with the underlying records | When the editorial explanation is reviewed |
| What must be checked? | Filters, units, states and permissions | Sources, context, claims and navigation |
Design the source and uncertainty into the page
A number needs a label, unit, period, geography and source. A missing record must not look like zero. A filter state needs to remain visible when someone copies a chart or reads it on a phone. When a story selects a subset, state that selection and offer a path to inspect the fuller dataset where permission allows.
Shiimain is a public prototype that begins with territory and then reveals geographic evidence and dashboard views. Janus Observatory uses a cinematic opening, atlas and explanatory layers for speculative futures. They show different ways to move between narrative and exploration. Neither case certifies the data for a new client or establishes measured reader understanding.
Scope the data pipeline as well as the interface
Ask who collects, cleans, approves and refreshes the data; what happens when an update fails; and which records are public. A dashboard with live filters may need authentication, access rules and monitoring. A published story may need editorial review, citations and a stable snapshot so the explanation still matches the figures readers see. Those are different delivery and maintenance responsibilities.
A useful first milestone combines one representative dataset, a view, its source note and the interaction that answers the main question. Test it with a person who did not design the chart. If they cannot state the unit, filter or limitation, improve the information design before adding more visual effects.
Sources & editorial notes
Sources
- Shiimain public prototypeshiimain.vercel.app
- Janus Observatory public experiencejanus-observatory.nicocerond.workers.dev
How we built this guide
EvidenceThis guide combines 2 linked sources, Dardo’s analysis and clearly labeled project examples.
LimitsVisible design and content decisions are not presented as measured ranking, traffic or conversion gains.
Read our editorial policy