NumPy Cheatsheet for Pandas Users (Beginner-Friendly)

Importing NumPy

import numpy as np

Creating NumPy Arrays

From Python Lists

arr = np.array([1, 2, 3, 4, 5])

From Pandas Series or DataFrame

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

# From DataFrame
df = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})
arr = df.to_numpy()

Basic Array Operations

Element-wise Operations

arr1 = np.array([1, 2, 3])
arr2 = np.array([4, 5, 6])

# Addition
result = arr1 + arr2

# Multiplication
result = arr1 * arr2

# Division
result = arr1 / arr2

Simple Mathematical Functions

# Square root
sqrt_arr = np.sqrt(arr)

# Exponential
exp_arr = np.exp(arr)

# Absolute value
abs_arr = np.abs(arr)

Statistical Operations

Basic Statistics

# Mean
mean = np.mean(arr)

# Median
median = np.median(arr)

# Standard deviation
std = np.std(arr)

Min, Max, and Sum

# Minimum
min_val = np.min(arr)

# Maximum
max_val = np.max(arr)

# Sum
total = np.sum(arr)

Array Manipulation

Reshaping

arr = np.array([1, 2, 3, 4, 5, 6])
reshaped = arr.reshape(2, 3)

Transposing

transposed = arr.T

Flattening

flattened = arr.flatten()

Random Number Generation

Random Sampling

# Generate 5 random numbers between 0 and 1
random_uniform = np.random.rand(5)

# Generate 5 random integers between 1 and 10
random_integers = np.random.randint(1, 11, 5)

Setting Random Seed

np.random.seed(42)  # For reproducibility

Working with Missing Data

Handling NaN Values

# Check for NaN
np.isnan(arr)

# Replace NaN with a value
np.nan_to_num(arr, nan=0.0)

Useful NumPy Functions for Pandas Users

unique() and value_counts()

# Get unique values
unique_values = np.unique(arr)

# Get value counts (similar to pandas value_counts())
values, counts = np.unique(arr, return_counts=True)

where()

# Similar to pandas' where, but returns an array
result = np.where(condition, x, y)

concatenate()

# Concatenate arrays (similar to pd.concat())
concatenated = np.concatenate([arr1, arr2, arr3])

When to Use NumPy with Pandas

Performance: For large datasets, NumPy operations can be faster than pandas.
Memory efficiency: NumPy arrays use less memory than pandas objects.
Specific mathematical operations: Some mathematical operations are more straightforward in NumPy.
Interfacing with other libraries: Many scientific Python libraries use NumPy arrays.

Remember, while these NumPy operations are useful, many have direct equivalents in pandas that work on Series and DataFrames. Always consider whether you can perform the operation directly in pandas before converting to NumPy arrays.