Data, Information and Knowledge
Everything in data analytics starts with a simple distinction that students routinely get wrong in exams: data is not information, and information is not knowledge.
The DIKW Pyramid
The classic model is the DIKW hierarchy — Data → Information → Knowledge → Wisdom. Each level adds context and value to the level below it.
Definitions
| Term | Definition | Key question it answers | Example |
|---|---|---|---|
| Data | Raw, unorganised facts, figures, symbols or observations with no meaning on their own | What? | 98, 76, 45, 88 |
| Information | Data that has been processed, organised and given context so it becomes meaningful | Who / What / Where / When? | "Average marks in Section A = 76.75" |
| Knowledge | Information combined with experience, interpretation and understood relationships; usable for decisions | How? | "Sections with average below 50 always have low attendance" |
| Wisdom | Applied knowledge with judgement about what should be done | Why / Should? | "Enforce a minimum attendance policy before exams" |
A Worked Progression
# ---------- DATA: raw values, no meaning ----------
values = [1200, 1450, 980, 1600, 2100]
# ---------- INFORMATION: add context + processing ----------
# These are daily sales (in rupees) for Mon-Fri at a campus canteen
days = ["Mon", "Tue", "Wed", "Thu", "Fri"]
sales = dict(zip(days, values))
print(sales)
# {'Mon': 1200, 'Tue': 1450, 'Wed': 980, 'Thu': 1600, 'Fri': 2100}
print("Total week sales:", sum(values)) # Total week sales: 7330
print("Average per day:", sum(values) / len(values)) # Average per day: 1466.0
# ---------- KNOWLEDGE: pattern noticed across many weeks ----------
# "Friday sales are consistently the highest; Wednesday is the weekly low."
# ---------- WISDOM: decision informed by that knowledge ----------
# "Staff two extra people on Friday; run a discount combo on Wednesday."Characteristics of Good Information
For data to become genuinely useful information, it must satisfy several quality attributes — a favourite short-answer question:
| Attribute | Meaning |
|---|---|
| Accuracy | Free from errors; correctly represents reality |
| Completeness | No essential values missing |
| Relevance | Actually related to the decision being made |
| Timeliness | Available when it is needed (stale data misleads) |
| Consistency | Same value everywhere it appears; no contradictions |
| Reliability | Comes from a trustworthy, verifiable source |
| Accessibility | Retrievable in a usable format by those who need it |
| Conciseness | Free from noise and irrelevant detail |
Data vs Information — Exam Comparison Table
| Basis | Data | Information |
|---|---|---|
| Meaning | Raw, unprocessed facts | Processed, organised data |
| Dependency | Independent (does not depend on information) | Depends on data |
| Context | Has no context | Always has context |
| Decision support | Cannot support decisions directly | Directly supports decisions |
| Unit of measure | Bits and bytes | Meaningful units (rupees, %, averages) |
| Example | 45, 90, "Delhi" | "Delhi branch scored 45% in Q1, 90% in Q2" |
Where Analytics Fits
Data Analytics is the discipline of applying processes and techniques to raw data in order to move it up this pyramid — extracting information, discovering knowledge, and ultimately supporting better decisions.
The remaining lessons of Unit 1 walk through exactly these arrows — how data is collected, sampled, cleaned, preprocessed and transformed so that analysis is even possible.