NumPy-Basic

A Comprehensive Guide to NumPy Fundamentals

From zero to hero: A structured learning path for mastering NumPy

Getting Started • Learning Path • Examples • Contributing • License

📋 Overview

This repository provides a comprehensive, step-by-step guide to mastering NumPy, the fundamental package for scientific computing in Python. Each phase builds systematically on previous knowledge, with practical examples and clear explanations.

🎯 Who Is This For?

• 🔰 Beginners looking to build a solid foundation in NumPy
• 🚀 Intermediate users wanting to deepen their understanding of advanced features
• 🎓 Students preparing for data science, machine learning, or AI coursework
• 💼 Professionals transitioning to roles requiring numerical computation skills

⚡ Quick Start

For those familiar with Python environments, get started immediately:

git clone https://github.com/Sourabh-Kumar04/Numpy-Basic.git
cd Numpy-Basic
python -m venv venv && source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt
jupyter notebook

Begin with Phase_1/01_phase_1.ipynb and progress through each phase sequentially.

📚 Learning Path

🔍 Phase-by-Phase Progress

Phase	Status	Topics	Key Concepts
🧩 Phase 1	✅ Complete	NumPy Fundamentals	Arrays vs Lists, Creating Arrays, Data Types, Basic Operations
📍 Phase 2	✅ Complete	Data Manipulation	Indexing, Slicing, Sorting, Boolean Masks, Fancy Indexing
🔄 Phase 3	✅ Complete	Array Transformation	Reshaping, Stacking, Splitting, Broadcasting Rules
🧮 Phase 4	🚧 In Progress	Advanced Topics	Vector/Matrix Operations, Trigonometric Functions, Statistics, File Operations

📁 Repository Structure

Numpy-Basic/
├── LICENSE                         # Apache 2.0 License
├── README.md                       # Project documentation
├── main.py                         # Example runner script
├── pyproject.toml                  # Project dependencies and configuration
├── uv.lock                         # Dependency lock file (for uv users)
├── Phase_1/
│   └── 01_phase_1.ipynb            # NumPy Basics: Arrays, Creation, Types
├── Phase_2/
│   └── 01_phase_2.ipynb            # Data Access: Indexing, Slicing, Filtering
├── Phase_3/
│   └── 01_phase_3.ipynb            # Data Transformation: Reshaping, Stacking, Broadcasting
└── Phase_4/
    ├── 01_phase_4.ipynb            # Advanced: Math Operations, Statistics, Visualization
    ├── array1.npy                  # Sample data file for practice
    ├── array2.npy                  # Sample data file for practice
    ├── array3.npy                  # Sample data file for practice
    └── numpy_logo.npy              # NumPy logo encoded as an array

🛠️ Installation & Setup

Prerequisites

• Python 3.8+ installed
• Git (for cloning the repository)
• Basic familiarity with Python programming

Step-by-Step Setup

1. Clone the repository

git clone https://github.com/Sourabh-Kumar04/Numpy-Basic.git
cd Numpy-Basic

2. Set up a virtual environment (Choose your preferred method)

# Option 1: Standard venv
python -m venv venv
source venv/bin/activate  # On macOS/Linux
# OR
venv\Scripts\activate     # On Windows

# Option 2: Using uv (faster alternative)
uv venv
uv activate

3. Install dependencies

# Option 1: Using pip
pip install -r requirements.txt

# Option 2: Using uv with pyproject.toml
uv pip install -e .

4. Launch Jupyter Notebook

jupyter notebook

📖 What You'll Learn

🧩 Phase 1: NumPy Fundamentals

• Why NumPy over standard Python lists?
  • Performance benchmarks showing speed differences
  • Memory efficiency comparisons
  • Vectorized operations
• Creating arrays from different sources
  • From Python lists
  • Using built-in functions: zeros(), ones(), arange(), linspace()
  • Random number generation
• Understanding array data types and properties

📍 Phase 2: Data Manipulation

• Accessing array elements
  • Basic indexing vs fancy indexing
  • Difference between views and copies
• Slicing multi-dimensional arrays
• Advanced selection with boolean masks
  • Filtering data with conditions
  • Combining multiple conditions
• Practical comparison between np.where() and boolean indexing
• Sorting arrays and finding unique values

🔄 Phase 3: Array Transformation

• Inspecting array properties
  • Shape, size, dimensions, data type
• Reshaping arrays
  • reshape(), ravel(), flatten()
  • Adding/removing dimensions with newaxis and squeeze()
• Combining arrays
  • Vertical stacking with vstack()
  • Horizontal stacking with hstack()
  • General stacking with concatenate()
• Broadcasting rules and compatibility
  • When operations work between arrays of different shapes
  • Common broadcasting errors and how to fix them

🧮 Phase 4: Applications & Advanced Features

• Vector, matrix, and tensor operations
  • Dot products, cross products, matrix multiplication
  • Linear algebra operations
• Comprehensive angle function reference
  • Trigonometric functions (sin, cos, tan)
  • Inverse trigonometric functions (arcsin, arccos, arctan2)
• Statistical functions for data analysis
  • Measures of central tendency
  • Measures of dispersion
  • Percentiles and quantiles
• Working with NumPy's native file formats
  • .npy for single arrays
  • .npz for multiple arrays
• Data visualization with matplotlib

📊 Code Examples

Performance Comparison: Lists vs. NumPy Arrays

import numpy as np
import time

# Python list operation
start = time.time()
python_list = list(range(1000000))
python_list = [x * 2 for x in python_list]
list_time = time.time() - start

# NumPy array operation
start = time.time()
numpy_array = np.arange(1000000)
numpy_array = numpy_array * 2
numpy_time = time.time() - start

print(f"Python list processing time: {list_time:.5f} seconds")
print(f"NumPy array processing time: {numpy_time:.5f} seconds")
print(f"NumPy is {list_time/numpy_time:.1f}x faster!")

Output

Python list processing time: 0.12345 seconds
NumPy array processing time: 0.00567 seconds
NumPy is 21.8x faster!

Fancy Indexing & Masking

import numpy as np

# Create sample data
data = np.random.randint(0, 100, size=(5, 5))
print("Original data:")
print(data)

# Boolean masking (values greater than 50)
mask = data > 50
filtered_data = data[mask]
print("\nValues greater than 50:")
print(filtered_data)

# Using np.where() for conditional values
result = np.where(data > 50, data * 2, data)
print("\nValues > 50 doubled, others unchanged:")
print(result)

Visualization with Dark Mode

import matplotlib.pyplot as plt
import numpy as np

# Set dark style
plt.style.use('dark_background')

# Generate data
x = np.linspace(0, 2 * np.pi, 100)
y1 = np.sin(x)
y2 = np.cos(x)

# Create plot
plt.figure(figsize=(10, 6))
plt.plot(x, y1, label='sin(x)', color='cyan', linewidth=2)
plt.plot(x, y2, label='cos(x)', color='magenta', linewidth=2)
plt.title("Trigonometric Functions", fontsize=16)
plt.xlabel("x (radians)", fontsize=12)
plt.ylabel("Amplitude", fontsize=12)
plt.legend(fontsize=12)
plt.grid(True, alpha=0.3)
plt.tight_layout()
plt.show()

📈 Statistical Functions Reference

Function	Description	Example
np.mean()	Arithmetic mean	np.mean(arr, axis=0)
np.median()	Median value	np.median(arr)
np.std()	Standard deviation	np.std(arr, ddof=1)
np.var()	Variance	np.var(arr)
np.min()	Minimum value	np.min(arr, axis=1)
np.max()	Maximum value	np.max(arr)
np.percentile()	nth percentile	np.percentile(arr, 75)
np.quantile()	nth quantile	np.quantile(arr, [0.25, 0.5, 0.75])
np.corrcoef()	Correlation coefficient	np.corrcoef(x, y)
np.cov()	Covariance matrix	np.cov(x, y)

💾 Working with .npy Files

import numpy as np

# Create sample array
array = np.random.normal(0, 1, size=(100, 100))

# Save to .npy file
np.save('sample_array.npy', array)

# Load from .npy file
loaded_array = np.load('sample_array.npy')

# Verify it's the same
print("Arrays are identical:", np.array_equal(array, loaded_array))

🤔 Common Questions & Answers

Why use NumPy instead of Python lists?

NumPy arrays are more efficient than Python lists for numerical operations because:

• They store data in contiguous memory blocks
• They leverage vectorized operations (SIMD instructions)
• They offer specialized numerical functions optimized in C
• They use less memory for the same amount of numerical data

What's the difference between a view and a copy?

• A view is just a different way to access the same data - changes to the view affect the original array
• A copy is a new array with the same values - changes to the copy don't affect the original
• Basic slicing typically returns views, while advanced indexing returns copies

What are broadcasting rules?

Broadcasting allows NumPy to perform operations on arrays of different shapes. The rules are:

• Arrays are compared from their trailing dimensions
• Dimensions with size 1 are stretched to match the other array
• Missing dimensions are treated as having size 1
• If dimensions are compatible, broadcasting proceeds

📚 Additional Resources

• Official NumPy Documentation
• NumPy Cheat Sheet (PDF)
• From Python to NumPy
• NumPy Tutorials
• 100 NumPy Exercises

🧠 How to Contribute

Contributions to improve this repository are welcome! Here's how you can help:

1. Fork the repository
2. Create a branch for your feature or fix
3. Commit your changes with descriptive messages
4. Push to your branch
5. Submit a Pull Request

Commit Message Convention

git commit -m "✨ Added Phase_3: Array reshaping and broadcasting examples"

Emoji	Description
✨	New features or content
🐛	Bug fixes
📝	Documentation updates
🔧	Configuration changes
🧹	Code cleanup
🎨	Style improvements

💬 Community & Support

• GitHub Discussions: Open a discussion
• Issue Tracker: Report bugs or request features

📄 License

This project is licensed under the Apache 2.0 License - see the LICENSE file for details.

📊 Citation

If you use this repository in your research or educational materials, please cite it as:

@misc{kumar2025numpybasic,
  author = {Kumar, Sourabh},
  title = {NumPy-Basic: Comprehensive Guide to NumPy Fundamentals},
  year = {2025},
  publisher = {GitHub},
  url = {https://github.com/Sourabh-Kumar04/Numpy-Basic},
  howpublished = {\url{https://github.com/Sourabh-Kumar04/Numpy-Basic}},
}

🌐 Connect & Support

• Author: Sourabh Kumar
• GitHub: @Sourabh-Kumar04
• LinkedIn: linkedin.com/in/sourabh-kumar04

---

If you find this repository helpful, please consider starring it! ⭐️

"NumPy doesn't just compute numbers—it transforms how we think about data." 🧮