Cross-tabulation with crosstab() in Data Analysis Python

Introduction

Cross-tabulation helps us see how two or more groups relate by counting how often their values appear together. It is like making a simple table to compare categories.

You want to compare how many people in different age groups prefer different types of fruits.

You want to check the relationship between gender and favorite sport in a survey.

You want to see how many customers from different cities bought different product categories.

You want to analyze how different education levels relate to job types in a dataset.

Syntax

Data Analysis Python

import pandas as pd

# Create a cross-tabulation table
pd.crosstab(index=data['Column1'], columns=data['Column2'], margins=False, normalize=False)

index is the row variable (categories for rows).

columns is the column variable (categories for columns).

Examples

Basic example showing counts of color and shape combinations.

Data Analysis Python

import pandas as pd

# Example data
colors = ['Red', 'Blue', 'Red', 'Green', 'Blue']
shapes = ['Circle', 'Square', 'Square', 'Circle', 'Circle']

# Cross-tabulation
pd.crosstab(index=colors, columns=shapes)

Edge case: empty data returns an empty DataFrame.

Data Analysis Python

import pandas as pd

# Empty data
colors = []
shapes = []

# Cross-tabulation on empty lists
pd.crosstab(index=colors, columns=shapes)

Edge case: single element returns a 1x1 table with count 1.

Data Analysis Python

import pandas as pd

# Single element data
colors = ['Red']
shapes = ['Circle']

# Cross-tabulation with one element
pd.crosstab(index=colors, columns=shapes)

Edge case: one unique column category, counts all in one column.

Data Analysis Python

import pandas as pd

# Data with one category in columns
colors = ['Red', 'Red', 'Red']
shapes = ['Circle', 'Circle', 'Circle']

# Cross-tabulation
pd.crosstab(index=colors, columns=shapes)

Sample Program

This program shows how many males and females prefer each fruit. It prints the original data and then the cross-tabulation table.

Data Analysis Python

import pandas as pd

# Sample data: survey of favorite fruit by gender
survey_data = {
    'Gender': ['Male', 'Female', 'Female', 'Male', 'Female', 'Male', 'Male'],
    'FavoriteFruit': ['Apple', 'Banana', 'Apple', 'Banana', 'Apple', 'Apple', 'Banana']
}

# Create DataFrame
survey_df = pd.DataFrame(survey_data)

print('Original DataFrame:')
print(survey_df)

# Create cross-tabulation table
fruit_gender_crosstab = pd.crosstab(index=survey_df['Gender'], columns=survey_df['FavoriteFruit'])

print('\nCross-tabulation of Gender and Favorite Fruit:')
print(fruit_gender_crosstab)

Important Notes

Time complexity is O(n), where n is the number of rows in the data.

Space complexity depends on the number of unique categories in the index and columns.

Common mistake: forgetting to pass the correct columns as index and columns, leading to confusing tables.

Use crosstab() when you want to count occurrences between categories. For more complex statistics, consider pivot tables.

Summary

Cross-tabulation counts how often categories from two variables appear together.

Use pd.crosstab() with index and columns to create the table.

It helps compare groups easily and understand relationships in data.

Practice

1. What does the pd.crosstab() function do in data analysis?

easy

A. It sorts data in ascending order.

B. It counts how often categories from two variables appear together.

C. It calculates the average of numerical columns.

D. It removes missing values from the dataset.

Solution

1. Step 1: Understand the purpose of crosstab()

   The function pd.crosstab() is used to count occurrences of combinations of categories from two variables.

2. Step 2: Compare with other options

   Sorting, averaging, and removing missing values are different functions, not related to crosstab.

3. Final Answer:

   It counts how often categories from two variables appear together. -> Option B

4. Quick Check:

   Cross-tabulation = counts of category pairs [OK]

Hint: Remember: crosstab counts category pairs, not calculations [OK]

Common Mistakes:

• Thinking crosstab sorts data
• Confusing crosstab with aggregation functions
• Assuming crosstab cleans data

2. Which of the following is the correct syntax to create a cross-tabulation table using pandas?

easy

A. pd.crosstab(index=df['AgeGroup'], columns=df['Gender'])

B. pd.crosstab(df.Gender, columns='AgeGroup')

C. pd.crosstab(columns=df['Gender'], index=df['AgeGroup'])

D. pd.crosstab(df['Gender'], df['AgeGroup'])

Solution

1. Step 1: Recall the correct parameter order for crosstab()

   The function pd.crosstab() takes the first argument as the row variable and the second as the column variable, both as Series or arrays.

2. Step 2: Check each option

   pd.crosstab(df['Gender'], df['AgeGroup']) correctly passes two Series without naming parameters. pd.crosstab(index=df['AgeGroup'], columns=df['Gender']) uses named parameters with row and column variables reversed. pd.crosstab(df.Gender, columns='AgeGroup') passes a string literal instead of a Series for the columns parameter. pd.crosstab(columns=df['Gender'], index=df['AgeGroup']) swaps index and columns incorrectly.

3. Final Answer:

   pd.crosstab(df['Gender'], df['AgeGroup']) -> Option D

4. Quick Check:

   Use positional args: rows, columns [OK]

Hint: Use positional args: first is rows, second is columns [OK]

Common Mistakes:

• Swapping index and columns parameters
• Mixing positional and named arguments incorrectly
• Passing DataFrame instead of Series

3. Given the DataFrame:

import pandas as pd
data = {'Color': ['Red', 'Blue', 'Red', 'Green', 'Blue', 'Green'],
        'Shape': ['Circle', 'Square', 'Square', 'Circle', 'Circle', 'Square']}
df = pd.DataFrame(data)
print(pd.crosstab(df['Color'], df['Shape']))

What is the output?

medium

A. Shape Circle Square
Color
Blue 1 1
Green 1 1
Red 1 1

B. Shape Circle Square
Color
Blue 2 0
Green 0 2
Red 1 1

C. Shape Circle Square
Color
Blue 1 0
Green 1 1
Red 1 1

D. Shape Circle Square
Color
Blue 1 1
Green 1 0
Red 1 1

Solution

1. Step 1: Count occurrences of each Color-Shape pair

   From the data:
   - Blue: Circle(1), Square(1)
   - Green: Circle(1), Square(1)
   - Red: Circle(1), Square(1)

2. Step 2: Construct the crosstab table

   Each color has one Circle and one Square, so the table shows 1 for each pair.

3. Final Answer:

   Shape Circle Square Color Blue 1 1 Green 1 1 Red 1 1 -> Option A

4. Quick Check:

   Count pairs correctly = Shape Circle Square Color Blue 1 1 Green 1 1 Red 1 1 [OK]

Hint: Count each category pair carefully from data [OK]

Common Mistakes:

• Miscounting pairs
• Swapping rows and columns
• Ignoring repeated categories

4. What is wrong with this code snippet?

import pandas as pd
data = {'Team': ['A', 'B', 'A', 'B'], 'Result': ['Win', 'Lose', 'Lose', 'Win']}
df = pd.DataFrame(data)
ct = pd.crosstab(index=df['Team'], columns='Result')
print(ct)

medium

A. The DataFrame is missing a column named 'Result'.

B. The index parameter should be a list, not a Series.

C. The columns parameter should be a Series, not a string.

D. The crosstab function requires numeric data only.

Solution

1. Step 1: Check the columns argument in crosstab()

   The columns argument must be a Series or array representing categories, not a string literal.

2. Step 2: Identify the error in the code

   Here, columns='Result' is a string, not df['Result']. This causes crosstab to treat it incorrectly.

3. Final Answer:

   The columns parameter should be a Series, not a string. -> Option C

4. Quick Check:

   Columns must be Series, not string [OK]

Hint: Pass Series for index and columns, not strings [OK]

Common Mistakes:

• Passing column name as string instead of Series
• Assuming crosstab needs numeric data
• Confusing index and columns parameters

5. You have a DataFrame with columns 'Department' and 'Satisfaction' (values: 'High', 'Medium', 'Low'). You want to find how many employees in each department have each satisfaction level, but only show departments with at least 3 employees total. Which code correctly does this?

hard

A. ct = pd.crosstab(df['Department'], df['Satisfaction'])
ct = ct[ct.sum(axis=1) >= 3]

B. ct = pd.crosstab(index=df['Satisfaction'], columns=df['Department'])
ct = ct[ct.sum(axis=0) >= 3]

C. ct = pd.crosstab(df['Department'], df['Satisfaction'])
ct = ct[ct.sum(axis=0) >= 3]

D. ct = pd.crosstab(index=df['Department'], columns=df['Satisfaction'])
ct = ct[ct.sum(axis=1) < 3]

Solution

1. Step 1: Create crosstab with Department as rows and Satisfaction as columns

   Use pd.crosstab(df['Department'], df['Satisfaction']) to count satisfaction levels per department.

2. Step 2: Filter departments with total employees >= 3

   Sum counts across columns (axis=1) for each department, then keep rows where sum >= 3.

3. Final Answer:

   ct = pd.crosstab(df['Department'], df['Satisfaction']) ct = ct[ct.sum(axis=1) >= 3] -> Option A

4. Quick Check:

   Filter rows by sum(axis=1) >= 3 [OK]

Hint: Sum rows (axis=1) to filter departments by total count [OK]

Common Mistakes:

• Filtering columns instead of rows
• Swapping index and columns in crosstab
• Using wrong axis in sum()