Chi-squared test in Data Analysis Python - Mini Project: Build & Apply

Chi-squared Test for Independence

📖 Scenario: You work in a marketing team. You want to check if customer preference for two different products depends on their age group. You have collected data on how many customers in each age group prefer each product.

🎯 Goal: Build a Python program that uses a chi-squared test to check if product preference depends on age group.

📋 What You'll Learn

Create a contingency table as a dictionary of dictionaries with exact counts

Create a variable for the significance level

Use scipy.stats.chi2_contingency to perform the chi-squared test

Print the chi-squared statistic and p-value

💡 Why This Matters

🌍 Real World

Chi-squared tests are used in marketing, medicine, and social sciences to check if two categorical variables are related.

💼 Career

Data analysts and scientists use chi-squared tests to analyze survey data and customer preferences to make informed decisions.

1

Create the contingency table data

Create a dictionary called preference_data with keys as age groups: '18-25', '26-35', and '36-45'. Each key should map to another dictionary with keys 'Product A' and 'Product B' and these exact counts: '18-25': {'Product A': 30, 'Product B': 20}, '26-35': {'Product A': 25, 'Product B': 25}, '36-45': {'Product A': 20, 'Product B': 30}.

Data Analysis Python

# Create the preference_data dictionary with exact counts
# Your code here

Hint

Think of preference_data as a table with age groups as rows and product preferences as columns.

2

Set the significance level

Create a variable called alpha and set it to 0.05 to represent the significance level for the chi-squared test.

Data Analysis Python

preference_data = {
    '18-25': {'Product A': 30, 'Product B': 20},
    '26-35': {'Product A': 25, 'Product B': 25},
    '36-45': {'Product A': 20, 'Product B': 30}
}
# Set the significance level alpha
# Your code here

Hint

The significance level alpha is usually set to 0.05 for many tests.

3

Perform the chi-squared test

Import chi2_contingency from scipy.stats. Convert preference_data into a 2D list called table with rows for each age group and columns for 'Product A' and 'Product B' in that order. Use chi2_contingency(table) to get the chi-squared statistic and p-value. Store them in variables chi2 and p.

Data Analysis Python

preference_data = {
    '18-25': {'Product A': 30, 'Product B': 20},
    '26-35': {'Product A': 25, 'Product B': 25},
    '36-45': {'Product A': 20, 'Product B': 30}
}

alpha = 0.05

# Import chi2_contingency and create the table list
# Perform the chi-squared test and store results in chi2 and p
# Your code here

Hint

Remember to import chi2_contingency before using it. The table must be a list of lists with counts in the right order.

4

Print the chi-squared test results

Print the chi-squared statistic with the text Chi-squared statistic: followed by the value of chi2. Then print the p-value with the text p-value: followed by the value of p. Use print statements.

Data Analysis Python

from scipy.stats import chi2_contingency

preference_data = {
    '18-25': {'Product A': 30, 'Product B': 20},
    '26-35': {'Product A': 25, 'Product B': 25},
    '36-45': {'Product A': 20, 'Product B': 30}
}

alpha = 0.05

table = [
    [preference_data['18-25']['Product A'], preference_data['18-25']['Product B']],
    [preference_data['26-35']['Product A'], preference_data['26-35']['Product B']],
    [preference_data['36-45']['Product A'], preference_data['36-45']['Product B']]
]

chi2, p, dof, expected = chi2_contingency(table)

# Print the chi-squared statistic and p-value
# Your code here

Hint

Use print(f"Chi-squared statistic: {chi2}") and similarly for p-value.