EDS 217 Cheatsheet

Introduction

This cheatsheet provides a quick reference for common operations on Pandas Series. It’s designed for beginning data science students who are just starting to work with Pandas.

Importing Pandas

Always start by importing pandas:

Code

import pandas as pd

Creating a Series

From a list

Code

data = [1, 2, 3, 4, 5]
s = pd.Series(data)
print(s)

0    1
1    2
2    3
3    4
4    5
dtype: int64

From a dictionary

Code

data = {'a': 0., 'b': 1., 'c': 2.}
s = pd.Series(data)
print(s)

a    0.0
b    1.0
c    2.0
dtype: float64

With custom index

Code

s = pd.Series([1, 2, 3, 4, 5], index=['a', 'b', 'c', 'd', 'e'])
print(s)

a    1
b    2
c    3
d    4
e    5
dtype: int64

Basic Series Information

Code

# Display the first few elements
print(s.head())

# Get basic information about the Series
print(s.info())

# Get summary statistics
print(s.describe())

# Get index
print(s.index)

# Get values
print(s.values)

a    1
b    2
c    3
d    4
e    5
dtype: int64
<class 'pandas.core.series.Series'>
Index: 5 entries, a to e
Series name: None
Non-Null Count  Dtype
--------------  -----
5 non-null      int64
dtypes: int64(1)
memory usage: 80.0+ bytes
None
count    5.000000
mean     3.000000
std      1.581139
min      1.000000
25%      2.000000
50%      3.000000
75%      4.000000
max      5.000000
dtype: float64
Index(['a', 'b', 'c', 'd', 'e'], dtype='object')
[1 2 3 4 5]

Selecting Data

By label

Code

# Select a single element
element = s['a']

# Select multiple elements
subset = s[['a', 'c', 'e']]

By position

Code

# Select by integer index (direct selection is being deprecated)
first_element = s.iloc[0]

# Select a slice by index (still okay):
subset = s[1:4]

# Select a slice (using iloc)
subset = s.iloc[1:4]

By condition

Code

# Select elements greater than 2
greater_than_two = s[s > 2]

Basic Data Manipulation

Updating values

Code

s['a'] = 10

Removing elements

Code

s=s.drop(labels=['a'])

Adding elements to a Series

Code

another_series = pd.Series(
    [3, 4, 5], 
    index=['f', 'g', 'h']
)
s = pd.concat([s, another_series])

The pd.concat() command (short for concatenate) is now the preferred method for extending series.

Important

The pd.concat() command takes a list of pd.Series objects to concatenate. This means you must create a pd.Series of new values to extend an existing pd.Series.

Updating elements based on their value using mask

Code

print(s)
s = s.mask(s > 2, s * 2)
print(s)

b    2
c    3
d    4
e    5
f    3
g    4
h    5
dtype: int64
b     2
c     6
d     8
e    10
f     6
g     8
h    10
dtype: int64

This line uses the greater than (>) logical operator within the mask() function to update the series. It will double the values in series where the condition s > 5 is True, while leaving other values unchanged.

Replacing elements based on their value using where

Code

print(s)
s = s.where(s < 8, 12)
print(s)

b     2
c     6
d     8
e    10
f     6
g     8
h    10
dtype: int64
b     2
c     6
d    12
e    12
f     6
g    12
h    12
dtype: int64

This line of code will update the values in series where condition is False (i.e. where s is not less than 8), replacing them with 12. The values where condition is True will remain unchanged.

Applying functions

Applying a newly-defined function

Code

def squared(x):
    ''' Square the value of x '''
    return x**2

# Apply a function to each element
s_squared = s.apply(squared)

Applying a lambda (temporary) function

Code

s_squared = s.apply(lambda x: x**2)

Lambda functions allow for doing transformations with temporary functions instead of having to define functions seperately. They are good for quick, 1-off transformations.

Handling missing values

Code

# Drop missing values
s_cleaned = s.dropna()

# Fill missing values
s_filled = s.fillna(0)  # Fills with 0

Basic Calculations

Code

# Calculate mean
mean_value = s.mean()

# Calculate sum
sum_value = s.sum()

# Calculate maximum
max_value = s.max()

Sorting

Code

print(s)
# Sort by values
s_sorted = s.sort_values()

# Sort by index
s_sorted_by_index = s.sort_index()

b     2
c     6
d    12
e    12
f     6
g    12
h    12
dtype: int64

Reindexing

Code

print(f"Original Series:\n{s}\n", sep='')
new_index = ['a', 'c', 'b', 'f', 'e', 'd', 'g', 'h', 'i', 'b']
s_reindexed = s.reindex(new_index)
print(f"Re-indexed series:\n{s_reindexed}")

Original Series:
b     2
c     6
d    12
e    12
f     6
g    12
h    12
dtype: int64

Re-indexed series:
a     NaN
c     6.0
b     2.0
f     6.0
e    12.0
d    12.0
g    12.0
h    12.0
i     NaN
b     2.0
dtype: float64

the reindex() command takes an ordered list that specifies the indicies that should be used to make a new pd.Series object. The list of indicies supplied to reindex() must have some indicies in common with the existing index. Indicies that do not appear in the existing Series will be set to NaN in the new Series. Repetition of indicies is allowed.

Combining Series

Code

s1 = pd.Series([1, 2, 3], index=['a', 'b', 'c'])
s2 = pd.Series([4, 5, 6], index=['d', 'e', 'f'])
s_combined = pd.concat([s1, s2])

Converting to Other Data Types

Code

# To list
list_data = s.tolist()

# To dictionary
dict_data = s.to_dict()

# To DataFrame
df = s.to_frame()

Further Learning

For more advanced operations and in-depth explanations, check out these resources:

Pandas Official Documentation
10 Minutes to Pandas
Python for Data Analysis by Wes McKinney

Remember, practice is key! Try these operations with different datasets to become more comfortable with Pandas Series.