NumPy

One 60-second card per skill, in learning order. 9 of 9 cards ready.

Topic 1

Array basics

Creating arrays · Shape and data type · Indexing and slicing

Open the lesson Practise this topic

60-second card

Creating arrays

A NumPy array is a grid of numbers of one type, built for fast maths.

np.array([1, 2, 3]) turns a list into an array.
np.zeros(3), np.ones((2, 3)) and np.arange(0, 10, 2) make arrays quickly.
np.linspace(0, 1, 5) gives 5 evenly spaced numbers.

import numpy as np
print(np.arange(0, 10, 2).tolist(), np.zeros(3).tolist())

Prints

[0, 2, 4, 6, 8] [0.0, 0.0, 0.0]

Remember: import numpy as np, then build arrays with np.something().

Common mistake: Expecting np.arange(0, 10, 2) to include 10. Like range, it stops before the end.

60-second card

Shape and data type

shape tells you the size of each dimension; dtype tells you the type of every value.

A 2×3 array has shape (2, 3): 2 rows, 3 columns.
ndim is the number of dimensions; size is the total count of values.
Mixing ints and floats makes everything float.

import numpy as np
a = np.array([[1, 2, 3], [4, 5, 6]])
print(a.shape, a.ndim, a.size, a.dtype)

Prints

(2, 3) 2 6 int64

Remember: Check .shape whenever something looks wrong.

Common mistake: Reading shape (2, 3) as 2 columns. It's always rows first.

60-second card

Indexing and slicing

Pick values with [row, column], and ranges with start:stop.

a[0, 2] is row 0, column 2.
a[:, 1] is every row of column 1.
a[1:3] takes positions 1 and 2; the stop is left out.

import numpy as np
a = np.array([[1, 2, 3], [4, 5, 6]])
print(a[0, 2], a[:, 1].tolist(), a[1, 1:].tolist())

Prints

3 [2, 5] [5, 6]

Remember: Rows first, then columns, separated by a comma.

Common mistake: Changing a slice and being surprised the original changed. Slices share data; use .copy().

Topic 2

Maths on arrays

Maths on whole arrays · Sum, mean, max and more · Broadcasting

Open the lesson Practise this topic

60-second card

Maths on whole arrays

Maths on an array works on every value at once, with no loop needed.

a * 2 doubles every value.
a + b adds matching positions of two arrays of the same shape.
np.sqrt(a) and np.round(a, 1) work on every value too.

import numpy as np
a = np.array([1, 2, 3])
b = np.array([10, 20, 30])
print((a * 2).tolist(), (a + b).tolist(), int(a @ b))

Prints

[2, 4, 6] [11, 22, 33] 140

Remember: No for loops: let NumPy do the whole array.

Common mistake: Expecting a * b to be matrix multiplication. It multiplies position by position; use @ for matrices.

60-second card

Sum, mean, max and more

Aggregations squash an array into one number, or one number per row or column.

sum(), mean(), max(), min() and std() are built in.
axis=0 works down each column; axis=1 works across each row.
argmax() gives the position of the largest value.

import numpy as np
a = np.array([[1, 2], [3, 4]])
print(a.sum(), a.sum(axis=0).tolist(), a.sum(axis=1).tolist())

Prints

10 [4, 6] [3, 7]

Remember: axis=0 → one result per column. axis=1 → one per row.

Common mistake: Mixing up the axes. Think: axis=0 collapses the rows.

60-second card

Broadcasting

Broadcasting lets NumPy combine arrays of different shapes by stretching the smaller one.

A single number is stretched to every position: a + 1.
A row of shape (3,) is added to each row of a (2, 3) array.
Shapes must match from the right, or be 1.

import numpy as np
a = np.array([[1, 2, 3], [4, 5, 6]])
print((a + np.array([10, 20, 30])).tolist())

Prints

[[11, 22, 33], [14, 25, 36]]

Remember: Line the shapes up from the right; each pair must be equal or 1.

Common mistake: Adding shapes (2, 3) and (2,). Their right ends (3 and 2) don't match, so it fails.

Topic 3

Reshaping and filtering

Reshaping · Filtering with conditions · Joining arrays

Open the lesson Practise this topic

60-second card

Reshaping

reshape changes an array's shape without changing its values.

np.arange(6).reshape(2, 3) makes 2 rows of 3.
Use -1 to let NumPy work out one size: reshape(3, -1).
flatten() or ravel() turns it back into one row.

import numpy as np
a = np.arange(6).reshape(3, -1)
print(a.shape, a.tolist())

Prints

(3, 2) [[0, 1], [2, 3], [4, 5]]

Remember: The total number of values must stay the same.

Common mistake: Reshaping 6 values into (4, 2). 4 × 2 = 8, so it fails.

60-second card

Filtering with conditions

A condition on an array gives True/False for each value. Use it to pick only the matching values.

a > 50 gives a mask like [False, True, True].
a[a > 50] keeps the values where the mask is True.
Combine with & (and) and | (or), with brackets: (a > 50) & (a < 90).

import numpy as np
marks = np.array([45, 72, 88, 91, 60])
print(marks[(marks > 50) & (marks < 90)].tolist())

Prints

[72, 88, 60]

Remember: Use & and |, not 'and' and 'or', with arrays.

Common mistake: Leaving out the brackets: a > 50 & a < 90 runs & first and goes wrong.

60-second card

Joining arrays

Join arrays end to end with concatenate, or stack them as rows or columns.

np.concatenate([a, b]) joins along an existing axis.
np.vstack stacks as rows; np.hstack joins side by side.
The other dimensions must match.

import numpy as np
a = np.array([1, 2])
b = np.array([3, 4])
print(np.concatenate([a, b]).tolist(), np.vstack([a, b]).tolist())

Prints

[1, 2, 3, 4] [[1, 2], [3, 4]]

Remember: vstack = on top. hstack = side by side.

Common mistake: Stacking arrays whose other dimension differs, like rows of length 3 and 4.