PYTHON PANDAS & NUMPY BASIC

📊 Python Pandas & NumPy – Complete Core Notes

Data Science Primitives: N-Dimensional Arrays, DataFrames, File I/O, Slicing & Data Manipulation ka reference guide.

1️⃣ NumPy Kya Hai?

Definition: NumPy (Numerical Python) Python ki core library hai jiska use multi-dimensional arrays (ndarray) aur high-performance mathematical computations ke liye kiya jata hai.

Python Lists vs NumPy Arrays: Python Lists alag-alag memory locations par pointers store karti hain, jabki NumPy arrays contiguous memory use karte hain aur C me implement hote hain. Is wajah se NumPy 50x se 100x fast execution provide karta hai.

• Vectorized Operations: Bina explicit for loops likhe pure array par mathematical calculations ek sath chal sakti hain.
• Broadcasting: Alag-alag shape ke arrays ke beech arithmetic calculations bina memory duplicate kiye handle karta hai.
• Scientific Foundation: Pandas, SciPy, Scikit-Learn aur TensorFlow jaise frameworks NumPy arrays ke upar hi build hain.

---

2️⃣ NumPy Array Operations

import numpy as np

# 1D & 2D Array Creation
arr = np.array([1, 2, 3, 4])
matrix = np.array([[1, 2], [3, 4]])

# Vectorized Arithmetic (No loops needed)
print(arr + 2)       # Output: [3 4 5 6]
print(arr * 3)       # Output: [3 6 9 12]

# Useful Array Generators
zeros = np.zeros((2, 3))       # 2x3 Matrix filled with 0.0
ranged = np.arange(1, 10, 2)   # [1, 3, 5, 7, 9]
lin = np.linspace(0, 1, 5)     # 0 aur 1 ke beech 5 barabar spaced values

---

3️⃣ Pandas Kya Hai?

Definition: Pandas (Python Data Analysis Library) structured, tabular data (Excel sheets, SQL tables, CSVs) ko clean, transform aur analyze karne ke liye primary data analysis library hai.

Data Structure	Dimensions	Real-World Analogy
Series	1-Dimensional	Excel sheet ka single column ya labelled 1D array.
DataFrame	2-Dimensional	Puri Excel spreadsheet jisme Rows aur Columns dono indexed hote hain.

---

4️⃣ DataFrame Creation & Structure

import pandas as pd

# Dictionary se DataFrame create karna
data = {
    'Name': ['Rahul', 'Anita', 'Ravi', 'Pooja'],
    'Age': [20, 22, 19, 24],
    'Score': [85, 92, 78, 95]
}

df = pd.DataFrame(data)
print(df)

Standard Output Representation:

    Name  Age  Score
0  Rahul   20     85
1  Anita   22     92
2   Ravi   19     78
3  Pooja   24     95

Interactive DataFrame Filter & Analysis Simulator

Name Age Score
Rahul 20 85
Anita 22 92
Ravi 19 78
Pooja 24 95

---

5️⃣ Reading & Writing Data Files

External datasets import aur export karne ke sabse useful methods:

File Format	Reading Command	Writing / Export Command
CSV	pd.read_csv('data.csv')	df.to_csv('out.csv', index=False)
Excel	pd.read_excel('data.xlsx')	df.to_excel('out.xlsx', index=False)
JSON	pd.read_json('data.json')	df.to_json('out.json')
SQL Table	pd.read_sql(query, connection)	df.to_sql('table', connection)

---

6️⃣ Data Selection & Filtering (loc vs iloc)

# 1. Single Column Access
names = df['Name']

# 2. loc (Label-based indexing: index name ya conditional logic)
senior_students = df.loc[df['Age'] > 20, ['Name', 'Score']]

# 3. iloc (Integer-based indexing: purely 0-indexed integer position)
first_row = df.iloc[0]        # Pehli row
first_two_rows = df.iloc[0:2] # Pehli 2 rows aur sabhi columns
specific_val = df.iloc[1, 0]  # Row 1, Column 0 ('Anita')

💡 Indexing Rule: iloc sirf pure integer values leta hai (jaise array indexing), jabki loc actual labels, column ke naam aur boolean conditional masks handle karta hai.

---

7️⃣ Common Data Wrangling Operations

# Column Addition & Dropping
df['Passed'] = df['Score'] >= 80          # Boolean calculation se naya column
df.drop('Passed', axis=1, inplace=True)    # Column delete karna (axis=1 means columns)

# Handling Missing / Null Values
df.dropna()                                # Null rows ko drop karna
df.fillna(0)                               # Null values ko 0 se replace karna

# Sorting & Aggregations
df.sort_values(by='Score', ascending=False)
mean_age = df['Age'].mean()
grouped = df.groupby('Age')['Score'].mean()