Array Shapes

What Is Array Shape?

The shape of a NumPy array describes the size of the array along each dimension.

For example:

import numpy as np

numbers = np.array([10, 20, 30, 40, 50])

print(numbers.shape)

(5,)

The result (5,) means the array contains 5 elements in one dimension.

Shape of a 1D Array

Consider this array:

numbers = np.array([10, 20, 30, 40, 50])

Visually, it looks like this:

10   20   30   40   50

It contains 5 values, so:

print(numbers.shape)

(5,)

This is a one-dimensional array.

Shape of a 2D Array

Now consider an array containing rows and columns:

data = np.array([
    [10, 20, 30],
    [40, 50, 60]
])

print(data)

[[10 20 30]
 [40 50 60]]

Let's count:

Rows    = 2
Columns = 3

Therefore the shape is:

print(data.shape)

(2, 3)

Read this as: 2 rows and 3 columns.

How to Read a Shape

This is one of the most important things to understand.

shape = (2, 3)

For a 2D array:

(rows, columns)

(2, 3)
 ↑   ↑
 │   └── 3 columns
 └────── 2 rows

So:

(2, 3)
→ 2 rows
→ 3 columns

More Shape Examples

Using the shape Property

NumPy arrays have a built-in property called shape.

import numpy as np

data = np.array([
    [10, 20, 30],
    [40, 50, 60]
])

print(data.shape)

(2, 3)

Notice that we use:

data.shape

not:

data.shape()

shape is a property, not a function.

Shape and Dimensions

Shape also helps us understand how many dimensions an array has.

For example:

numbers = np.array([10, 20, 30, 40])

print(numbers.shape)
print(numbers.ndim)

(4,)
1

Here:

shape = (4,)
ndim  = 1

So this is a one-dimensional array containing four elements.

Checking a 2D Array

data = np.array([
    [10, 20, 30],
    [40, 50, 60]
])

print(data.shape)
print(data.ndim)

(2, 3)
2

This tells us:

shape → (2, 3)
        2 rows
        3 columns

ndim → 2 dimensions

Three-Dimensional Arrays

NumPy can also work with three-dimensional arrays.

data = np.array([
    [
        [1, 2],
        [3, 4]
    ],
    [
        [5, 6],
        [7, 8]
    ]
])

print(data.shape)
print(data.ndim)

(2, 2, 2)
3

The shape is:

(2, 2, 2)

You can think of this as:

2 blocks
×
2 rows
×
2 columns

Three-dimensional structures become important when working with things such as image data and neural networks.

Why Shape Matters in AI

Machine Learning and Deep Learning models expect data in specific shapes.

For example, imagine we have data for 100 students, and each student has 3 features:

Age
Height
Weight

The data could have a shape of:

(100, 3)

This means:

100 → students
3   → features for each student

So the array could look conceptually like:

[
    [age, height, weight],
    [age, height, weight],
    [age, height, weight],
    ...
]

Understanding this becomes critical when preparing datasets for Machine Learning models.

Array Shape and Images

Images are another good example of why shape matters.

Suppose a grayscale image has:

28 rows
28 columns

Its shape can be:

(28, 28)

A color image usually has an additional dimension for the color channels.

For example:

(224, 224, 3)

This can represent:

224 → image height
224 → image width
3   → color channels

This is why understanding array shapes is essential for computer vision and Deep Learning.

Shape vs Size

Do not confuse shape with size.

data = np.array([
    [10, 20, 30],
    [40, 50, 60]
])

print(data.shape)
print(data.size)

(2, 3)
6

The shape tells us the structure:

(2, 3)
→ 2 rows
→ 3 columns

The size tells us the total number of elements:

2 × 3 = 6

Complete Example

Let's put everything together:

import numpy as np

students = np.array([
    [20, 170, 65],
    [22, 175, 70],
    [25, 180, 80],
    [21, 168, 60]
])

print("Data:")
print(students)

print("Shape:", students.shape)
print("Dimensions:", students.ndim)
print("Total elements:", students.size)

Data:
[[ 20 170  65]
 [ 22 175  70]
 [ 25 180  80]
 [ 21 168  60]]

Shape: (4, 3)
Dimensions: 2
Total elements: 12

Now we can clearly understand the result:

Shape
(4, 3)

4 → students
3 → features

Dimensions
2 → rows and columns

Size
12 → 4 × 3

A Simple Rule to Remember

When you see a NumPy shape, read it from left to right.

(5,)
→ 5 elements

(2, 3)
→ 2 rows × 3 columns

(4, 5)
→ 4 rows × 5 columns

(2, 3, 4)
→ 2 × 3 × 4

The number of values inside the shape tells you the number of dimensions.

(5,)       → 1 dimension
(2, 3)     → 2 dimensions
(2, 3, 4)  → 3 dimensions