Pandas Series Cheat Sheet

Pandas Series - One-dimensional labeled array capable of holding any data type

NUMPY ARRAY VS PANDAS SERIES

Key Differences

When to Use What

Code Comparison

Index Alignment Example

# NumPy - No alignment, just position-based
arr1 = np.array([1, 2, 3])
arr2 = np.array([4, 5, 6])
arr1 + arr2  # [5, 7, 9] - adds by position

# Pandas - Automatic index alignment
s1 = pd.Series([1, 2, 3], index=['a', 'b', 'c'])
s2 = pd.Series([4, 5, 6], index=['b', 'c', 'd'])
s1 + s2  # a: NaN, b: 6, c: 8, d: NaN - aligns by label

Conversion Between NumPy and Pandas

Performance Considerations

# NumPy - Fastest for pure numeric operations
arr = np.array([1, 2, 3, 4, 5])
arr * 2  # Very fast

# Pandas - Slightly slower but more features
s = pd.Series([1, 2, 3, 4, 5])
s * 2  # Fast, plus handles NaN, alignment, etc.

# For large datasets with numeric operations only → NumPy
# For data analysis with labels/missing data → Pandas

QUICK REFERENCE

Import & Creation

Basic Properties

INDEXING & SELECTION

Basic Indexing

Advanced Selection

OPERATIONS

Arithmetic Operations

Comparison Operations

Logical Operations

AGGREGATION & STATISTICS

Basic Statistics

Advanced Statistics

Cumulative Operations

MISSING DATA

Detection

Handling

SORTING & RANKING

Sorting

Ranking

STRING OPERATIONS

Basic String Methods

String Extraction

DATETIME OPERATIONS

Datetime Properties

Datetime Methods

TRANSFORMATION

Apply & Map

Binning & Categorization

Type Conversion

COMBINING SERIES

Concatenation

Set Operations

RESHAPING

Duplicates

Reindexing

GROUPING & AGGREGATION

GroupBy Operations

WINDOW FUNCTIONS

Rolling Windows

Expanding Windows

Exponential Weighted

PLOTTING

Basic Plots

I/O OPERATIONS

Reading

Writing

COMMON PATTERNS

Creating Series

# From list
s = pd.Series([1, 2, 3, 4, 5])

# With custom index
s = pd.Series([1, 2, 3], index=['a', 'b', 'c'])

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

# With name
s = pd.Series([1, 2, 3], name='my_series')

# From range
s = pd.Series(range(10))

Filtering

# Single condition
s[s > 5]

# Multiple conditions
s[(s > 2) & (s < 8)]

# Using isin
s[s.isin([1, 3, 5])]

# Using between
s[s.between(2, 8)]

Handling Missing Data

# Drop NaN
s.dropna()

# Fill with value
s.fillna(0)

# Forward fill
s.fillna(method='ffill')

# Interpolate
s.interpolate()

String Operations

# Convert to lowercase
s.str.lower()

# Check contains
s.str.contains('pattern')

# Extract numbers
s.str.extract(r'(\d+)')

# Split and expand
s.str.split(',', expand=True)

Datetime Operations

# Convert to datetime
s = pd.to_datetime(s)

# Extract components
s.dt.year
s.dt.month
s.dt.day

# Format
s.dt.strftime('%Y-%m-%d')

Aggregation

# Basic stats
s.mean()
s.median()
s.std()

# Value counts
s.value_counts()

# Describe
s.describe()

# Custom aggregation
s.agg(['mean', 'std', 'min', 'max'])

PERFORMANCE TIPS

1. Use vectorized operations instead of loops

   # Good: s * 2
   # Bad: s.apply(lambda x: x * 2)
   

2. Use .loc[] and .iloc[] explicitly for clarity and performance

3. Chain operations for readability

   s.dropna().sort_values().head(10)
   

4. Use .copy() to avoid SettingWithCopyWarning

   s_new = s[s > 5].copy()
   

5. Use categorical dtype for repeated string values

   s = s.astype('category')
   

QUICK TIPS

• Index alignment: Operations automatically align on index
• Broadcasting: Scalar operations broadcast to all elements
• NaN handling: Most operations skip NaN by default
• Method chaining: Most methods return Series for chaining
• Inplace operations: Use inplace=True to modify in place
• Copy vs view: Use .copy() when you need independent data

Print Settings: Landscape orientation, narrow margins, fit to page
Pro Tip: Laminate for durable desk reference!

Last Updated: January 2026