NumPy Detailed Tutorial (Beginner to Advanced)

NumPy is the foundation of Python data and AI workflows. If you understand NumPy deeply, Pandas, Matplotlib, and ML libraries become much easier.

1) Install and Setup

pip install numpy

import numpy as np

2) Core Concept: ndarray

a = np.array([10, 20, 30, 40])
print(a.shape)   # (4,)
print(a.dtype)   # usually int64

Important attributes:

• shape: dimensions
• ndim: number of dimensions
• dtype: data type
• size: total elements

3) Creating Arrays

np.zeros((2, 3))
np.ones((3, 3))
np.arange(0, 10, 2)
np.linspace(0, 1, 5)
np.eye(3)

Use linspace when you need exact number of points, not step size.

4) Indexing, Slicing, and Filtering

x = np.array([5, 8, 12, 3, 9, 15])
print(x[2])       # 12
print(x[1:4])     # [ 8 12  3]
print(x[x > 8])   # [12  9 15]

2D example:

m = np.array([[1, 2, 3], [4, 5, 6]])
print(m[0, 2])    # 3
print(m[:, 1])    # column 1 -> [2 5]

5) Broadcasting (Most Important Intermediate Skill)

Broadcasting allows arithmetic between arrays of compatible shapes.

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

Output:

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

Rule summary:

• Compare shapes from right to left
• Dimensions are compatible if equal, or one of them is 1

6) Vectorization (Avoid Python Loops)

# slow style (loop)
arr = np.arange(1_000_000)
out = [x * 2 for x in arr]

# fast NumPy style
fast = arr * 2

Vectorized operations are typically much faster and cleaner.

7) Aggregations and Statistics

data = np.array([3, 5, 8, 10, 12])
print(data.sum())
print(data.mean())
print(data.std())
print(data.min(), data.max())

Axis-based example:

mat = np.array([[1, 2, 3], [4, 5, 6]])
print(mat.sum(axis=0))  # column sums
print(mat.sum(axis=1))  # row sums

8) Reshape, Stack, and Split

a = np.arange(12)
b = a.reshape(3, 4)
c = np.vstack([b, b])
d = np.hstack([b, b])

Be careful: reshape requires same total element count.

9) Linear Algebra Basics

A = np.array([[2, 1], [1, 3]])
B = np.array([[1, 0], [0, 1]])

print(A @ B)                 # matrix multiplication
print(np.linalg.det(A))      # determinant
print(np.linalg.inv(A))      # inverse
print(np.linalg.eigvals(A))  # eigenvalues

10) Random Sampling for Simulation

rng = np.random.default_rng(42)
print(rng.integers(0, 10, size=5))
print(rng.normal(loc=0, scale=1, size=1000))

Use default_rng instead of old global random APIs for cleaner reproducibility.

11) Practical Mini Project

Build a simple sales analytics array pipeline:

1. Generate random daily sales for 3 products over 30 days
2. Compute total and average sales by product
3. Find top 5 highest sales days overall
4. Normalize each product series with z-score

12) Common Mistakes

• Mixing Python lists and arrays unintentionally
• Forgetting axis in reductions
• Confusing element-wise * with matrix multiplication @
• Using loops where vectorization is possible

Next Step

After NumPy, move to:

• Pandas Detailed Tutorial: Data Cleaning, Analysis, and Real Workflows