Matrix As Data Structure

Definition

A matrix can be used as a data structure.

It stores values in rows and columns.

Example:

[70 180 30]
[65 140 25]
[72 200 40]

Each row could represent one person.

Each column could represent one feature.

Matrix as data structure = organized storage of many related values.

---

Rows As Items

Rows can represent data items.

Example:

[70 180 30]
[65 140 25]
[72 200 40]

Rows:

person 1 = [70, 180, 30]
person 2 = [65, 140, 25]
person 3 = [72, 200, 40]

Each row is a vector.

A matrix can store many vectors together.

---

Columns As Features

Columns can represent features.

Example:

[70 180 30]
[65 140 25]
[72 200 40]

Columns could mean:

column 1 = height
column 2 = weight
column 3 = age

Every row must use the same column meaning.

Columns make component meaning consistent across data items.

---

Matrix Shape As Dataset Shape

If a matrix stores data, its shape can describe the dataset.

Example:

3 x 4 matrix

could mean:

3 examples
4 features per example

Another example:

100 x 20 matrix

could mean:

100 examples
20 features per example

Matrix shape can describe how many items and features are stored.

---

Consistency Requirement

A matrix is useful only if rows and columns are consistent.

Good:

[height weight age]
[height weight age]
[height weight age]

Bad:

[height weight age]
[age height weight]
[weight age height]

The bad matrix mixes feature positions.

That breaks comparison and transformation.

Matrix data depends on consistent column meaning.

---

Matrix vs Table

A table can contain different kinds of values.

Example:

name, age, city

A matrix, in this chapter, is numerical.

Example:

[70 180 30]
[65 140 25]

The numerical structure matters because matrices can be transformed mathematically.

A matrix is like a numerical table with mathematical structure.

---

Why This Matters

Matrices let computation work with many vectors at once.

Instead of processing one vector:

[70, 180, 30]

we can organize many:

[70 180 30]
[65 140 25]
[72 200 40]

This is the basic shape of many datasets.

Matrices let vector data scale from one item to many items.

---

Final Summary

• A matrix can store structured numerical data.
• Rows can represent items.
• Columns can represent features.
• Each row can be treated as a vector.
• Column meaning must stay consistent.
• Matrix shape can describe dataset shape.
• Matrices let computation handle many related vectors.