DIKW Pyramid
Definition
A hierarchical model of how raw data becomes meaningful action, ascending through four (sometimes five) levels: Data → Information → Knowledge → Wisdom (sometimes Understanding is inserted between Knowledge and Wisdom).
Origins traced to T.S. Eliot’s poem The Rock (1934), formalized by Russell Ackoff (1989) in From Data to Wisdom.
The pyramid
WISDOM = applied judgement, knowing when/why
▲
KNOWLEDGE = patterns, models, know-how
▲
INFORMATION = structured, contextualized data
▲
DATA = raw facts, signals, measurements
Why it matters for NOTQIN
The 5-layer model in the Operating Thesis (Physical → Signal → Event → KPI → Decision) is essentially DIKW applied to the factory floor:
| DIKW | Industrial equivalent (Thesis §4) |
|---|---|
| Data | Signal (PLC tag values) |
| Information | Event (machine_stopped, defect_detected) |
| Knowledge | KPI (OEE, MTBF, OTIF) |
| Wisdom | Decision (stop the line, escalate, dispatch maintenance) |
Most “AI for factories” tools live at the Information layer. NOTQIN’s wedge is making the Knowledge→Wisdom transition explicit and auditable.
Example
- Data:
temperature_sensor_03 = 84.2 - Information: Stamping press #3 ran at 84.2°C during shift 2 on 2026-05-24
- Knowledge: Press #3 typically runs at 72°C ± 4; >82°C predicts tool wear within 200 cycles
- Wisdom: Schedule preventive tool change at next break to avoid mid-shift line stop costing 165k MAD/min
Common mistakes
- Calling a dashboard “knowledge” when it only shows Information
- Confusing more Data with more Wisdom — there’s no automatic ascent without deliberate processing
- Treating DIKW as strictly hierarchical (some philosophers reject it; Zeleny added Enlightenment above Wisdom)
- Building a “knowledge base” that’s actually an information catalog