crosstab() for cross-tabulation in Pandas - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is the purpose of the crosstab() function in pandas?

The crosstab() function creates a table that shows the frequency (count) of combinations between two or more categorical variables. It helps to summarize and compare data easily.

beginner

How do you create a simple cross-tabulation between two columns col1 and col2 in a DataFrame df?

Use pd.crosstab(df['col1'], df['col2']). This counts how many times each combination of values from col1 and col2 appears.

intermediate

What does the normalize parameter do in crosstab()?

The normalize parameter changes counts into proportions or percentages. For example, normalize='index' shows row-wise proportions, and normalize='columns' shows column-wise proportions.

intermediate

Can crosstab() handle more than two variables? How?

Yes, by passing multiple arrays or columns as arguments. For example, pd.crosstab([df['col1'], df['col2']], df['col3']) creates a multi-index table showing counts for combinations of col1 and col2 against col3.

beginner

What is a real-life example where crosstab() is useful?

Imagine a survey with answers about gender and favorite fruit. crosstab() can show how many males and females prefer each fruit, helping to understand preferences by group.

What does pd.crosstab(df['A'], df['B']) return?

A A table counting occurrences of each combination of values in columns A and B

B A table showing the sum of values in columns A and B

C A table with the mean of columns A and B

D A table with unique values from columns A and B

crosstab() counts how often each pair of values appears in the two columns.

How do you get proportions instead of counts in crosstab()?

A Set <code>normalize='index'</code> or <code>normalize='columns'</code>

B Set <code>normalize='all'</code>

C Set <code>normalize=True</code>

D Set <code>normalize=False</code>

Use normalize='index' for row proportions or normalize='columns' for column proportions.

Which of these is a valid way to use crosstab() with three variables?

A <code>pd.crosstab(df['A'], df['B'], df['C'])</code>

B <code>pd.crosstab(df['A'])</code>

C <code>pd.crosstab(df['A'] + df['B'], df['C'])</code>

D <code>pd.crosstab([df['A'], df['B']], df['C'])</code>

Passing a list of columns as the first argument allows cross-tabulation on multiple variables.

What type of data is best suited for crosstab()?

A Numerical continuous data

B Categorical data

C Time series data

D Text data

crosstab() works best with categorical data to count combinations.

If you want to see how many customers bought product A or B by region, which function helps?

A groupby()

B pivot_table()

C crosstab()

D merge()

crosstab() quickly shows counts of combinations like product and region.

Explain how to use crosstab() to analyze the relationship between two categorical columns in a DataFrame.

Think about how to count how often each pair of categories appears.

Describe how the normalize parameter changes the output of crosstab() and why it might be useful.

Consider when percentages are easier to understand than raw counts.

Practice

1. What does the pandas.crosstab() function primarily do?

easy

A. Filters rows based on a condition

B. Sorts data in ascending order

C. Counts how often values from two columns appear together

D. Calculates the mean of numerical columns

Solution

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

   The crosstab() function creates a table showing frequency counts of combinations of values from two or more columns.

2. Step 2: Compare with other options

   Sorting, calculating mean, and filtering are different operations not done by crosstab().

3. Final Answer:

   Counts how often values from two columns appear together -> Option C

4. Quick Check:

   crosstab() = frequency count [OK]

Hint: Remember: crosstab counts pairs of values [OK]

Common Mistakes:

• Confusing crosstab with sorting or filtering
• Thinking crosstab calculates averages
• Assuming crosstab modifies original data

2. Which of the following is the correct syntax to create a crosstab of columns 'A' and 'B' from DataFrame df?

easy

A. pd.crosstab(df.A, df.B, axis=1)

B. pd.crosstab(df['A'], df['B'])

C. df.crosstab('A', 'B')

D. pd.crosstab(df['A', 'B'])

Solution

1. Step 1: Recall correct crosstab syntax

   The correct syntax is pd.crosstab(df['A'], df['B']) where you pass two series as arguments.

2. Step 2: Identify incorrect options

   pd.crosstab(df.A, df.B, axis=1) wrongly uses axis parameter; df.crosstab('A', 'B') tries to call crosstab as a DataFrame method which doesn't exist; pd.crosstab(df['A', 'B']) passes a tuple incorrectly.

3. Final Answer:

   pd.crosstab(df['A'], df['B']) -> Option B

4. Quick Check:

   pd.crosstab(series1, series2) = correct syntax [OK]

Hint: Use pd.crosstab(series1, series2) syntax [OK]

Common Mistakes:

• Calling crosstab as a DataFrame method
• Passing multiple columns as a tuple incorrectly
• Using wrong parameters like axis

3. Given the DataFrame:

import pandas as pd
data = {'Gender': ['M', 'F', 'F', 'M', 'F'], 'Preference': ['Tea', 'Coffee', 'Tea', 'Coffee', 'Tea']}
df = pd.DataFrame(data)

What is the output of pd.crosstab(df['Gender'], df['Preference'])?

medium

A. Preference Coffee Tea
Gender
F 1 2
M 1 1

B. Preference Coffee Tea
Gender
F 2 1
M 1 1

C. Preference Coffee Tea
Gender
F 1 1
M 2 1

D. Preference Coffee Tea
Gender
F 1 2
M 2 0

Solution

1. Step 1: Count occurrences for each Gender and Preference

   For 'F': Coffee appears once, Tea appears twice. For 'M': Coffee appears once, Tea appears once.

2. Step 2: Construct the crosstab table

   Rows are Gender (F, M), columns are Preference (Coffee, Tea), with counts filled accordingly.

3. Final Answer:

   Preference Coffee Tea Gender F 1 2 M 1 1 -> Option A

4. Quick Check:

   Count values by groups = Preference Coffee Tea Gender F 1 2 M 1 1 [OK]

Hint: Count each group pair carefully for crosstab [OK]

Common Mistakes:

• Mixing up row and column labels
• Counting values incorrectly
• Assuming missing pairs have counts > 0

4. What is wrong with this code snippet?

import pandas as pd
data = {'A': ['x', 'y', 'x'], 'B': ['p', 'q', 'p']}
df = pd.DataFrame(data)
result = pd.crosstab(df['A'], df['B'], margins=True, margin_name='Total')
print(result)

It raises an error. Why?

medium

A. The print statement is missing parentheses

B. You cannot use margins=True with crosstab

C. The DataFrame columns must be numeric for crosstab

D. The parameter name should be margins_name, not margin_name

Solution

1. Step 1: Check parameter names for crosstab()

   The correct parameter to name the totals row/column is margins_name, not margin_name.

2. Step 2: Identify the cause of error

   Using margin_name causes a TypeError because it's an unexpected keyword argument.

3. Final Answer:

   The parameter name should be margins_name, not margin_name -> Option D

4. Quick Check:

   Correct parameter is margins_name [OK]

Hint: Use margins_name, not margin_name for totals [OK]

Common Mistakes:

• Misspelling parameter names
• Thinking crosstab requires numeric columns
• Forgetting parentheses in print (Python 3)

5. You have a DataFrame df with columns 'City', 'Product', and 'Sales'. You want to create a crosstab showing total sales for each city-product pair. Which code correctly achieves this?

hard

A. pd.crosstab(df['City'], df['Product'], values=df['Sales'], aggfunc='sum')

B. pd.crosstab(df['City'], df['Product'], values='Sales', aggfunc='sum')

C. pd.crosstab(df['City'], df['Product'], aggfunc='sum')

D. pd.crosstab(df['City'], df['Product'], values=df['Sales'])

Solution

1. Step 1: Understand crosstab with aggregation

   To aggregate values like sales, use values= with the column and aggfunc= with the aggregation function.

2. Step 2: Check each option

   pd.crosstab(df['City'], df['Product'], values=df['Sales'], aggfunc='sum') correctly passes the sales column and sum function. pd.crosstab(df['City'], df['Product'], values='Sales', aggfunc='sum') passes a string instead of a series for values. pd.crosstab(df['City'], df['Product'], aggfunc='sum') misses values parameter. pd.crosstab(df['City'], df['Product'], values=df['Sales']) misses aggfunc, so it defaults to count.

3. Final Answer:

   pd.crosstab(df['City'], df['Product'], values=df['Sales'], aggfunc='sum') -> Option A

4. Quick Check:

   Use values= and aggfunc= for aggregation [OK]

Hint: Use values=column and aggfunc='sum' for totals [OK]

Common Mistakes:

• Passing column name as string instead of series
• Omitting aggfunc for aggregation
• Expecting crosstab to sum without values parameter