🐼 Pandas Handbook

Master data manipulation and analysis with DataFrames, Series, and powerful data operations

1. Introduction to Pandas

Pandas is the most popular Python library for data manipulation and analysis. It provides powerful, flexible data structures designed for working with structured (tabular, multidimensional, potentially heterogeneous) and time series data.

Why Pandas? Fast and efficient DataFrame objects, tools for reading/writing data, data alignment and missing data handling, reshaping and pivoting

Pandas is built on top of NumPy and is a key library for:

  • Data Cleaning - Handle missing data, duplicates, and inconsistencies
  • Data Transformation - Reshape, pivot, merge, and aggregate
  • Data Analysis - Statistical analysis and exploration
  • Time Series - Work with dates, times, and time-indexed data

2. Installation and Import

Bash 
# Install with pip
$ pip install pandas

# Or with Anaconda
$ conda install pandas
Python 
import pandas as pd
import numpy as np

# Check version
print(pd.__version__)  # 1.3.0 or newer

3. Series

A Series is a one-dimensional labeled array capable of holding any data type.

Python 
# Create Series from list
s = pd.Series([1, 3, 5, 7, 9])
print(s)
# 0    1
# 1    3
# 2    5
# 3    7
# 4    9

# Series with custom index
s = pd.Series([1, 3, 5, 7, 9], index=['a', 'b', 'c', 'd', 'e'])
print(s)
# a    1
# b    3
# c    5
# d    7
# e    9

# Create from dictionary
data = {'a': 1, 'b': 2, 'c': 3}
s = pd.Series(data)
print(s)

# Accessing elements
print(s['a'])     # 1
print(s[0])       # 1
print(s[['a', 'c']])  # Multiple elements

# Series operations
print(s + 10)     # Add 10 to all elements
print(s * 2)      # Multiply all by 2
print(s[s > 1])   # Filter values > 1

4. DataFrames

A DataFrame is a 2-dimensional labeled data structure with columns of potentially different types.

Creating DataFrames

Python 
# From dictionary
data = {
    'Name': ['Alice', 'Bob', 'Charlie', 'David'],
    'Age': [25, 30, 35, 40],
    'City': ['New York', 'London', 'Paris', 'Tokyo']
}
df = pd.DataFrame(data)
print(df)
#       Name  Age      City
# 0    Alice   25  New York
# 1      Bob   30    London
# 2  Charlie   35     Paris
# 3    David   40     Tokyo

# From list of lists
data = [
    ['Alice', 25, 'New York'],
    ['Bob', 30, 'London'],
    ['Charlie', 35, 'Paris']
]
df = pd.DataFrame(data, columns=['Name', 'Age', 'City'])

# From NumPy array
arr = np.array([[1, 2, 3], [4, 5, 6]])
df = pd.DataFrame(arr, columns=['A', 'B', 'C'])

# Basic info
print(df.shape)      # (4, 3) - rows, columns
print(df.columns)    # Column names
print(df.index)      # Row indices
print(df.dtypes)     # Data types
print(df.info())     # Summary info

5. Reading Data

Pandas can read data from various file formats.

Python 
# Read CSV
df = pd.read_csv('data.csv')

# Read with specific options
df = pd.read_csv('data.csv', 
                 sep=',',           # Delimiter
                 header=0,          # Row number to use as column names
                 index_col=0,       # Column to use as row labels
                 na_values=['NA', 'missing'])  # Values to treat as NaN

# Read Excel
df = pd.read_excel('data.xlsx', sheet_name='Sheet1')

# Read JSON
df = pd.read_json('data.json')

# Read from SQL database
import sqlite3
conn = sqlite3.connect('database.db')
df = pd.read_sql_query("SELECT * FROM table_name", conn)

# Read from URL
url = 'https://example.com/data.csv'
df = pd.read_csv(url)

6. Viewing and Inspecting Data

Python 
# View first/last rows
print(df.head())      # First 5 rows
print(df.head(10))    # First 10 rows
print(df.tail())      # Last 5 rows

# Quick statistics
print(df.describe())  # Summary statistics
print(df.info())      # DataFrame info

# Shape and size
print(df.shape)       # (rows, columns)
print(df.size)        # Total elements
print(len(df))        # Number of rows

# Check for missing values
print(df.isnull().sum())      # Count nulls per column
print(df.notnull().sum())     # Count non-nulls

# Unique values
print(df['City'].unique())     # Unique values in column
print(df['City'].nunique())    # Count unique values
print(df['City'].value_counts())  # Frequency of each value

# Column information
print(df.columns.tolist())     # List of column names
print(df.dtypes)               # Data types of columns

7. Selection and Indexing

Selecting Columns

Python 
# Single column (returns Series)
ages = df['Age']
print(ages)

# Multiple columns (returns DataFrame)
subset = df[['Name', 'Age']]
print(subset)

# Add new column
df['Salary'] = [50000, 60000, 70000, 80000]

# Delete column
df = df.drop('City', axis=1)  # axis=1 for columns
# or
del df['City']

Selecting Rows

Python 
# Select by position with iloc
print(df.iloc[0])         # First row
print(df.iloc[0:3])       # First 3 rows
print(df.iloc[0:3, 0:2])  # First 3 rows, first 2 columns

# Select by label with loc
df = df.set_index('Name')
print(df.loc['Alice'])            # Row with index 'Alice'
print(df.loc['Alice':'Charlie'])  # Rows from Alice to Charlie
print(df.loc['Alice', 'Age'])     # Specific cell

# Boolean indexing
print(df.iloc[[0, 2]])    # Rows 0 and 2

8. Filtering Data

Python 
# Single condition
df_filtered = df[df['Age'] > 30]
print(df_filtered)

# Multiple conditions (AND)
df_filtered = df[(df['Age'] > 25) & (df['Salary'] > 55000)]

# Multiple conditions (OR)
df_filtered = df[(df['Age'] < 30) | (df['City'] == 'Tokyo')]

# Using isin()
cities = ['New York', 'London']
df_filtered = df[df['City'].isin(cities)]

# Using query() method
df_filtered = df.query('Age > 30 and Salary < 75000')

# Filter by string methods
df_filtered = df[df['Name'].str.startswith('A')]
df_filtered = df[df['Name'].str.contains('li')]

# Not null filtering
df_filtered = df[df['Age'].notnull()]

9. Data Operations

Sorting

Python 
# Sort by column
df_sorted = df.sort_values('Age')
df_sorted = df.sort_values('Age', ascending=False)

# Sort by multiple columns
df_sorted = df.sort_values(['City', 'Age'])

# Sort index
df_sorted = df.sort_index()

Adding and Modifying Columns

Python 
# Add new column from calculation
df['Age_in_10_years'] = df['Age'] + 10

# Add column from condition
df['Senior'] = df['Age'] > 35

# Modify existing column
df['Age'] = df['Age'] + 1

# Rename columns
df = df.rename(columns={'Age': 'Years', 'City': 'Location'})

# Drop duplicates
df = df.drop_duplicates()
df = df.drop_duplicates(subset=['Name'])  # Based on specific column

String Operations

Python 
# Convert case
df['Name'] = df['Name'].str.upper()
df['Name'] = df['Name'].str.lower()
df['Name'] = df['Name'].str.title()

# String methods
df['Name'].str.len()              # Length of strings
df['Name'].str.strip()            # Remove whitespace
df['Name'].str.replace('A', 'a')  # Replace characters
df['Name'].str.split(' ')         # Split strings

# Extract patterns
df['City'].str.extract(r'([A-Z][a-z]+)')  # Regex extraction

10. GroupBy Operations

Group data and perform aggregate operations.

Python 
# Basic groupby
grouped = df.groupby('City')

# Aggregate functions
print(df.groupby('City')['Age'].mean())
print(df.groupby('City')['Salary'].sum())
print(df.groupby('City').size())  # Count per group

# Multiple aggregations
print(df.groupby('City').agg({
    'Age': ['mean', 'min', 'max'],
    'Salary': ['sum', 'mean']
}))

# Group by multiple columns
df.groupby(['City', 'Department']).mean()

# Apply custom function to groups
def age_range(x):
    return x.max() - x.min()

df.groupby('City')['Age'].apply(age_range)

# Iterate through groups
for name, group in df.groupby('City'):
    print(f"\n{name}:")
    print(group)

11. Merging and Joining

Concatenation

Python 
# Vertical concatenation (stack rows)
df1 = pd.DataFrame({'A': [1, 2], 'B': [3, 4]})
df2 = pd.DataFrame({'A': [5, 6], 'B': [7, 8]})
result = pd.concat([df1, df2], ignore_index=True)

# Horizontal concatenation (stack columns)
result = pd.concat([df1, df2], axis=1)

# Concatenate with keys
result = pd.concat([df1, df2], keys=['first', 'second'])

Merging (SQL-style joins)

Python 
df1 = pd.DataFrame({
    'key': ['A', 'B', 'C'],
    'value1': [1, 2, 3]
})

df2 = pd.DataFrame({
    'key': ['A', 'B', 'D'],
    'value2': [4, 5, 6]
})

# Inner join (default)
result = pd.merge(df1, df2, on='key')
# Only keeps rows where key exists in both

# Left join
result = pd.merge(df1, df2, on='key', how='left')
# Keeps all rows from df1

# Right join
result = pd.merge(df1, df2, on='key', how='right')
# Keeps all rows from df2

# Outer join
result = pd.merge(df1, df2, on='key', how='outer')
# Keeps all rows from both

# Merge on multiple columns
result = pd.merge(df1, df2, on=['key1', 'key2'])

# Merge with different column names
result = pd.merge(df1, df2, left_on='key1', right_on='key2')

12. Handling Missing Data

Python 
# Check for missing values
print(df.isnull())           # Boolean DataFrame
print(df.isnull().sum())     # Count per column
print(df.isnull().any())     # Any nulls per column

# Drop missing values
df_clean = df.dropna()                 # Drop rows with any null
df_clean = df.dropna(axis=1)           # Drop columns with any null
df_clean = df.dropna(how='all')        # Drop only if all values are null
df_clean = df.dropna(subset=['Age'])   # Drop based on specific column

# Fill missing values
df_filled = df.fillna(0)                      # Fill with 0
df_filled = df.fillna(df.mean())              # Fill with mean
df_filled = df.fillna(method='ffill')         # Forward fill
df_filled = df.fillna(method='bfill')         # Backward fill

# Fill specific column
df['Age'] = df['Age'].fillna(df['Age'].mean())

# Interpolate missing values
df['Age'] = df['Age'].interpolate()

# Replace specific values
df = df.replace(0, np.nan)                    # Replace 0 with NaN
df = df.replace([0, -1], np.nan)              # Replace multiple values

13. Apply Functions

Apply custom functions to DataFrames and Series.

Python 
# Apply to Series
def double(x):
    return x * 2

df['Age_doubled'] = df['Age'].apply(double)

# Lambda function
df['Age_squared'] = df['Age'].apply(lambda x: x ** 2)

# Apply to DataFrame rows (axis=1)
def calculate_bonus(row):
    if row['Age'] > 35:
        return row['Salary'] * 0.1
    else:
        return row['Salary'] * 0.05

df['Bonus'] = df.apply(calculate_bonus, axis=1)

# Apply to DataFrame columns (axis=0)
df_normalized = df[['Age', 'Salary']].apply(lambda x: (x - x.min()) / (x.max() - x.min()))

# Map values (for Series)
age_category = {
    25: 'Young',
    30: 'Mid',
    35: 'Senior',
    40: 'Senior'
}
df['Category'] = df['Age'].map(age_category)

# Replace with mapping
df['City'] = df['City'].replace({
    'New York': 'NY',
    'Los Angeles': 'LA'
})

# applymap (element-wise on entire DataFrame)
df_rounded = df[['Age', 'Salary']].applymap(lambda x: round(x, 2))

14. Time Series

Working with dates, times, and time-indexed data.

Creating Date Ranges

Python 
# Date range
dates = pd.date_range('2023-01-01', periods=10, freq='D')
print(dates)

# Different frequencies
dates = pd.date_range('2023-01-01', periods=12, freq='M')   # Month
dates = pd.date_range('2023-01-01', periods=52, freq='W')   # Week
dates = pd.date_range('2023-01-01', periods=24, freq='H')   # Hour

# Create DataFrame with date index
df = pd.DataFrame({
    'value': np.random.randn(10)
}, index=dates)

Date Operations

Python 
# Convert string to datetime
df['date'] = pd.to_datetime(df['date_string'])

# Parse with specific format
df['date'] = pd.to_datetime(df['date_string'], format='%Y-%m-%d')

# Extract date components
df['year'] = df['date'].dt.year
df['month'] = df['date'].dt.month
df['day'] = df['date'].dt.day
df['day_name'] = df['date'].dt.day_name()
df['week'] = df['date'].dt.isocalendar().week

# Filter by date
df_filtered = df[df['date'] > '2023-01-01']
df_filtered = df[(df['date'] >= '2023-01-01') & (df['date'] <= '2023-12-31')]

# Set date as index
df = df.set_index('date')

# Resample time series data
df_monthly = df.resample('M').mean()      # Monthly average
df_weekly = df.resample('W').sum()        # Weekly sum

# Rolling window calculations
df['rolling_mean'] = df['value'].rolling(window=7).mean()
df['rolling_sum'] = df['value'].rolling(window=7).sum()

# Shift data (for lag features)
df['previous_value'] = df['value'].shift(1)
df['next_value'] = df['value'].shift(-1)

15. Exporting Data

Save DataFrames to various file formats.

Python 
# Export to CSV
df.to_csv('output.csv', index=False)
df.to_csv('output.csv', index=False, sep=';')  # Custom separator

# Export to Excel
df.to_excel('output.xlsx', sheet_name='Data', index=False)

# Multiple sheets
with pd.ExcelWriter('output.xlsx') as writer:
    df1.to_excel(writer, sheet_name='Sheet1', index=False)
    df2.to_excel(writer, sheet_name='Sheet2', index=False)

# Export to JSON
df.to_json('output.json')
df.to_json('output.json', orient='records')  # List of records

# Export to SQL
from sqlalchemy import create_engine
engine = create_engine('sqlite:///database.db')
df.to_sql('table_name', engine, if_exists='replace', index=False)
# if_exists options: 'fail', 'replace', 'append'

# Export to HTML
df.to_html('output.html')

# Export to pickle (preserves data types)
df.to_pickle('output.pkl')

# Read pickle
df = pd.read_pickle('output.pkl')

Advanced Export Options

Python 
# Export specific columns
df[['Name', 'Age']].to_csv('subset.csv', index=False)

# Export with compression
df.to_csv('output.csv.gz', compression='gzip', index=False)

# Export with encoding
df.to_csv('output.csv', encoding='utf-8', index=False)

# Export to clipboard (for quick paste)
df.to_clipboard(index=False)

Bonus: Common Pandas Patterns

Chaining Operations

Python 
# Method chaining for clean code
result = (df
    .query('Age > 25')
    .groupby('City')['Salary']
    .mean()
    .sort_values(ascending=False)
    .head(5)
)

# Pipeline approach
result = (df
    .assign(Bonus=lambda x: x['Salary'] * 0.1)
    .query('Bonus > 5000')
    .sort_values('Bonus', ascending=False)
)

Pivot Tables

Python 
# Create pivot table
pivot = df.pivot_table(
    values='Salary',
    index='City',
    columns='Department',
    aggfunc='mean'
)

# Multiple aggregations
pivot = df.pivot_table(
    values='Salary',
    index='City',
    columns='Department',
    aggfunc=['mean', 'sum', 'count']
)

# Melt (unpivot)
melted = df.melt(
    id_vars=['Name'],
    value_vars=['Age', 'Salary'],
    var_name='Metric',
    value_name='Value'
)

Cross Tabulation

Python 
# Frequency table
crosstab = pd.crosstab(df['City'], df['Department'])

# With percentages
crosstab = pd.crosstab(
    df['City'], 
    df['Department'],
    normalize='index'  # Row percentages
)

# With margins (totals)
crosstab = pd.crosstab(
    df['City'], 
    df['Department'],
    margins=True
)

Window Functions

Python 
# Ranking
df['rank'] = df['Salary'].rank(ascending=False)
df['rank'] = df.groupby('City')['Salary'].rank(ascending=False)

# Cumulative operations
df['cumsum'] = df['Salary'].cumsum()
df['cummax'] = df['Salary'].cummax()

# Percent change
df['pct_change'] = df['Salary'].pct_change()

# Expanding window (cumulative statistics)
df['expanding_mean'] = df['Salary'].expanding().mean()

🎉 Congratulations!

You've mastered Pandas fundamentals! Ready to query databases with SQL?

Next: SQL → ← Back to NumPy 🏠 Home