Data as Representation of Facts

Data as Representation of Facts | Generated by AI

Home 2026.08

Question: Is this definition of data correct?

数据：对客观事物的性质、状态以及相互关系等进行记载的符号及其组合，通常是数字、文字、图像等，也可以是其他抽象的符号。

Answer:

Yes, this is basically correct. It is a classic definition from information science / database theory.

A more precise understanding:

Data = representation of facts.

The key points:

Data describes things

The “things” can be:
- objects: a person, a car, a product
- events: a transaction, a click, a temperature reading
- relationships: “Alice bought Product A”
Data itself is symbols, not meaning

Example:

2026-08-01

This is data.

Its meaning depends on context:

2026-08-01 = date
2026-08-01 = product ID
2026-08-01 = version number

The raw symbols become information after interpretation.

A common hierarchy:

Data
  |
  | interpretation
  v
Information
  |
  | analysis + experience
  v
Knowledge
  |
  | decision
  v
Wisdom

Example:

Raw data:

{
  "user_id": 123,
  "temperature": 39.2,
  "time": "08:00"
}

Information:

User 123 had a high temperature at 08:00.

Knowledge:

A temperature above 39°C may indicate fever.

Decision:

Recommend medical attention.

In computer science, we usually generalize data into:

1. Structured data

Fixed schema:

User(
    id INT,
    name VARCHAR,
    age INT
)

Examples:

databases
CSV
JSON with fixed fields

2. Semi-structured data

Has structure but flexible:

{
  "name": "Alice",
  "skills": ["Python", "AI"]
}

Examples:

JSON
XML
logs

3. Unstructured data

No fixed schema:

images
audio
video
documents
natural language

Modern AI is largely about learning representations from unstructured data:

text
 |
Tokenizer
 |
tokens
 |
embeddings
 |
vectors
 |
model learns patterns

For an LLM:

"The cat sits on the mat"

is converted into:

tokens:
[464, 3797, 286, 262, 2603]

↓

embedding vectors:

[0.12, -0.54, ...]

The model never directly sees “meaning”; it learns statistical patterns over numerical representations.

So your definition is correct. A more modern AI-oriented version would be:

Data is a symbolic representation of objects, events, attributes, and relationships that can be stored, transmitted, processed, and interpreted by humans or machines.

References:

Database System Concepts
Artificial Intelligence: A Modern Approach

Back Donate