2/17/25, 11:35 PM Walmart_Business_Case_Study.

ipynb - Colab

import gdown
!gdown 1lcjtmvtSjco6cnWaJvgJNTGjGJoKieh1

Downloading...
From: https://drive.google.com/uc?id=1lcjtmvtSjco6cnWaJvgJNTGjGJoKieh1
To: /content/walmart-data.csv
100% 23.0M/23.0M [00:00<00:00, 35.3MB/s]

# Importing required libraries -

# For reading & manipulating the data -

import pandas as pd
import numpy as np

# For visualizing the data -

import matplotlib.pyplot as plt
import seaborn as sns

# For statistical functions -

import scipy.stats as stats

# Loading the dataset -

df = pd.read_csv('walmart-data.csv')
df.sample(10)
df.head(10)
df.tail(10)

Show hidden output

# Shape of the dataset

df.shape
print('No. of rows:', df.shape[0])
print('No. of columns:', df.shape[1])

No. of rows: 550068

No. of columns: 10

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 550068 entries, 0 to 550067
Data columns (total 10 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 User_ID 550068 non-null int64
1 Product_ID 550068 non-null object
2 Gender 550068 non-null object
3 Age 550068 non-null object
4 Occupation 550068 non-null int64
5 City_Category 550068 non-null object
6 Stay_In_Current_City_Years 550068 non-null object
7 Marital_Status 550068 non-null int64
8 Product_Category 550068 non-null int64
9 Purchase 550068 non-null int64

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 1/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab
dtypes: int64(5), object(5)
memory usage: 42.0+ MB

keyboard_arrow_down Delete Null values and outliers

# Checking for null values:-
df.isnull().sum()
df.isna().sum()

User_ID 0

Product_ID 0

Gender 0

Age 0

Occupation 0

City_Category 0

Stay_In_Current_City_Years 0

Marital_Status 0

Product_Category 0

Purchase 0

dtype: int64

No null values found

# Checking for duplicates

duplicate_rows = df[df.duplicated()]
print(duplicate_rows.shape[0])

No duplicate rows found

sns.kdeplot(data=df, x='Purchase')
plt.show()

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 2/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

attrs = ['Gender', 'Age', 'Occupation', 'City_Category', 'Stay_In_Current_City_Years', 'Marital_Status']

sns.set_style("white")

fig, axs = plt.subplots(nrows=3, ncols=2, figsize=(20, 16))

fig.subplots_adjust(top=1.3)
count = 0
for row in range(3):
for col in range(2):
sns.boxplot(data=df, y='Purchase', x=attrs[count], hue=attrs[count], ax=axs[row, col], palette='Set3', legend=False)
axs[row, col].set_title(f"Purchase vs {attrs[count]}", pad=12, fontsize=13)
count += 1
plt.show()

Show hidden output

q1=df.Purchase.quantile(0.25)
q3=df.Purchase.quantile(0.75)
print(q1,q3)
IQR=q3-q1
outliers = df[((df.Purchase<(q1-1.5*IQR)) | (df.Purchase>(q3+1.5*IQR)))]
print("num outliers : ",len(outliers))
print("percent outliers : ",len(outliers)/len(df))

5823.0 12054.0
num outliers : 2677
percent outliers : 0.004866671029763593

df_clean=df.drop(df[ (df.Purchase > (q3+1.5*IQR)) | (df.Purchase < (q1-1.5*IQR)) ].index)

df_clean.info()

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 3/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

<class 'pandas.core.frame.DataFrame'>
Index: 547391 entries, 0 to 550067
Data columns (total 10 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 User_ID 547391 non-null int64
1 Product_ID 547391 non-null object
2 Gender 547391 non-null object
3 Age 547391 non-null object
4 Occupation 547391 non-null int64
5 City_Category 547391 non-null object
6 Stay_In_Current_City_Years 547391 non-null object
7 Marital_Status 547391 non-null int64
8 Product_Category 547391 non-null int64
9 Purchase 547391 non-null int64
dtypes: int64(5), object(5)
memory usage: 45.9+ MB

df=df_clean.reset_index()

keyboard_arrow_down Data Exploration

# Statistical summary of the dataset -
df.describe(include='all').T

Show hidden output

df['User_ID'].nunique()

5891

df['Product_ID'].nunique()

3631

df['Gender'].value_counts()/len(df)

count

Gender

M 0.752974

F 0.247026

dtype: float64

Categorical_rows = ['Gender','Age','Stay_In_Current_City_Years','Marital_Status','City_Category']
df[Categorical_rows].melt().groupby(['variable','value'])[['value']].count()*100/len(df)

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 4/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

value

variable value

Age 0-17 2.746117

18-25 18.146809

26-35 39.946035

36-45 19.987358

46-50 8.301561

51-55 6.976914

55+ 3.895205

City_Category A 26.861238

B 42.038324

C 31.100438

Gender F 24.702635

M 75.297365

Marital_Status 0 59.051391

1 40.948609

Stay_In_Current_City_Years 0 13.525250

1 35.229845

2 18.521313

3 17.319247

4+ 15.404345

1. Age Distribution The majority of users (around 40%) fall in the 26-35 age group, making it the largest customer segment. The 36-45 age
group follows, comprising 20% of users. Younger users (18-25) make up 18%, while older users (46+ years) collectively contribute around
19%. Very few users (~3%) are under the age of 17.

2. City Category Distribution The highest proportion of users (42%) come from City Category B. City C contributes 31% of users, while City A
accounts for 27%. This suggests that users from mid-sized cities (B & C) form the majority of the customer base.

3. Gender Distribution A significantly higher percentage of users (75%) are male, while only 25% are female. This indicates that the platform
or product is more popular among male customers.

4. Marital Status A larger proportion (59%) of users are single, compared to 41% who are married. This suggests that the product may appeal
more to unmarried individuals.

5. Stay Duration in Current City The majority of users (35%) have been staying in their current city for 1 year, followed by 18% for 2 years.
Around 17% have lived in their city for 3 years, and 15% for 4+ years. A relatively smaller proportion (13.5%) are newcomers (0 years). This
could indicate a mix of both stable and migrating users.

- Overall Summary

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 5/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab
The core customer base consists of males aged 26-35 from City Category B. A large proportion of users are single and have lived in their city
for 1-3 years. City B & C residents form the majority, while City A has the lowest representation.

fig, axs = plt.subplots(nrows=2, ncols=2, figsize=(16, 12))

sns.countplot(data=df, x='Gender', ax=axs[0,0])
sns.countplot(data=df, x='Age', ax=axs[0,1])
sns.countplot(data=df, x='City_Category', ax=axs[1,0])
sns.countplot(data=df, x='Marital_Status', ax=axs[1,1])
plt.show()

Show hidden output

sns.barplot(data=df, x='Age', y='Purchase',hue="Gender")

<Axes: xlabel='Age', ylabel='Purchase'>

sns.barplot(data=df, x='Age', y='Purchase',hue="Marital_Status")

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 6/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

<Axes: xlabel='Age', ylabel='Purchase'>

plt.figure(figsize=(10, 8))
sns.countplot(data=df, x='Product_Category')
plt.show()

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 7/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

plt.figure(figsize=(20, 8))
sns.countplot(data=df, x='Age',hue='Marital_Status')
plt.show()

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 8/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

As expected, there are no married individuals in the 0-17 age group.

Up to the age of 45, the number of unmarried individuals is higher than married ones.
After the age of 45, the number of married individuals exceeds the unmarried ones.

plt.figure(figsize=(20, 8))
sns.countplot(data=df, x='Product_Category',hue='Marital_Status')
plt.show()

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 9/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

keyboard_arrow_down How does gender affect the amount spent?

# Checking different metrics based on purchase by different genders -
df.groupby('Gender')['Purchase'].describe()

count mean std min 25% 50% 75% max

Gender

F 135220.0 8671.049039 4679.058483 12.0 5429.0 7906.0 11064.0 21398.0

M 412171.0 9367.724355 5009.234088 12.0 5852.0 8089.0 12247.0 21399.0

## run it multiple times

df.sample(500).groupby('Gender')['Purchase'].describe()

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 10/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

count mean std min 25% 50% 75% max

Gender

F 107.0 8739.476636 4092.670974 743.0 5996.0 8041.0 10551.0 20634.0

M 393.0 8877.592875 4569.020437 136.0 5938.0 7994.0 11729.0 20419.0

sns.displot(x='Purchase', bins=25, kde=True,hue='Gender', data=df )

<seaborn.axisgrid.FacetGrid at 0x7a9b40f3d890>

We can see that the distribution is close to normal.

sns.displot(x='Purchase', data=df, bins=25, hue='Gender')

plt.axvline(x=df['Purchase'].mean(), color='r')
plt.axvline(x=df[df['Gender']=='M']['Purchase'].mean(), color='b')
plt.axvline(x=df[df['Gender']=='F']['Purchase'].mean(), color='g')

plt.show()

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 11/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

The mean purchase amounts for males, females, and the overall dataset are quite similar, with females spending slightly less on average.

Let's dive deeper into the details.

keyboard_arrow_down Female Purchases

# Female Purchases
Female_data = df[df["Gender"]=="F"]
print("Female Purchase amount Mean:"+ str(Female_data["Purchase"].mean()))
print("Female Purchase amount SD:"+ str(Female_data["Purchase"].std()))

Female Purchase amount Mean:8671.049038603756

Female Purchase amount SD:4679.058483084425

keyboard_arrow_down Let us pick around 1,000 random samples of size 300 from the entire data set and calculate the
mean of each sample.

female_sample_means=[df[df['Gender']=='F'].sample(300, replace=True)['Purchase'].mean() for i in range(1000)]

sns.displot(female_sample_means,bins=30, kde=True )

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 12/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

<seaborn.axisgrid.FacetGrid at 0x7a9b42b2ce90>

females_sample_means_600=[df[df['Gender']=='F'].sample(600, replace=True)['Purchase'].mean() for i in range(1000)]

sns.displot(females_sample_means_600,bins=30, kde=True )

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 13/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

<seaborn.axisgrid.FacetGrid at 0x7a9b43d43450>

As the sample size increases, the sampling distribution becomes more normally distributed, with reduced deviation.

print("Sample distribution mean for females : ",sum(females_sample_means_600) / len(females_sample_means_600))

Sample distribution mean for females : 8675.96046499999

# Generate bootstrap sample means for females

female_sample_means = [df[df['Gender']=='F'].sample(1000, replace=True)['Purchase'].mean() for i in range(1000)]

# Compute the 95% Confidence Interval

lower_bound = np.percentile(female_sample_means, 2.5)
upper_bound = np.percentile(female_sample_means, 97.5)

# Visualization
plt.figure(figsize=(10, 6))
sns.histplot(female_sample_means, bins=30, kde=True, color="purple")

# Add vertical lines for confidence interval

plt.axvline(lower_bound, color='red', linestyle='dashed', label=f'Lower 95% CI: {lower_bound:.2f}')
plt.axvline(upper_bound, color='red', linestyle='dashed', label=f'Upper 95% CI: {upper_bound:.2f}')
plt.axvline(np.mean(female_sample_means), color='blue', linestyle='solid', label=f'Mean: {np.mean(female_sample_means):.2f}')

# Labels and title

plt.xlabel("Sample Mean of Purchase (Female)")
plt.ylabel("Frequency")
plt.title("Bootstrap Distribution of Female Purchase Means with 95% CI")

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 14/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab
plt.legend()
plt.show()

# Print confidence interval

print(f"95% Confidence Interval for Female Purchase Mean: ({lower_bound:.2f}, {upper_bound:.2f})")

95% Confidence Interval for Female Purchase Mean: (8393.90, 8977.83)

keyboard_arrow_down Male's Purchases

# Male Purchases
Male_data = df[df["Gender"]=="M"]
print("Male Purchase amount Mean:"+ str(Male_data["Purchase"].mean()))
print("Male Purchase amount SD:"+ str(Male_data["Purchase"].std()))

Male Purchase amount Mean:9367.724354697444

Male Purchase amount SD:5009.234087946683

male_sample_means=[df[df['Gender']=='M'].sample(300, replace=True)['Purchase'].mean() for i in range(1000)]

sns.displot(male_sample_means,bins=30, kde=True )

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 15/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

<seaborn.axisgrid.FacetGrid at 0x7a9b46d12610>

males_sample_means_600=[df[df['Gender']=='M'].sample(600, replace=True)['Purchase'].mean() for i in range(1000)]

sns.displot(males_sample_means_600,bins=30, kde=True )

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 16/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

<seaborn.axisgrid.FacetGrid at 0x7a9b47134c90>

print("Sample distribution mean for males : ",sum(male_sample_means) / len(male_sample_means))

Sample distribution mean for males : 9371.82929899999

# Generate bootstrap sample means for females

male_sample_means = [df[df['Gender']=='M'].sample(1000, replace=True)['Purchase'].mean() for i in range(1000)]

# Compute the 95% Confidence Interval

lower_bound = np.percentile(male_sample_means, 2.5)
upper_bound = np.percentile(male_sample_means, 97.5)

# Visualization
plt.figure(figsize=(10, 6))
sns.histplot(male_sample_means, bins=30, kde=True, color="purple")

# Add vertical lines for confidence interval

plt.axvline(lower_bound, color='red', linestyle='dashed', label=f'Lower 95% CI: {lower_bound:.2f}')
plt.axvline(upper_bound, color='red', linestyle='dashed', label=f'Upper 95% CI: {upper_bound:.2f}')
plt.axvline(np.mean(male_sample_means), color='blue', linestyle='solid', label=f'Mean: {np.mean(male_sample_means):.2f}')

# Labels and title

plt.xlabel("Sample Mean of Purchase (Male)")
plt.ylabel("Frequency")
plt.title("Bootstrap Distribution of Male Purchase Means with 95% CI")
plt.legend()
plt.show()

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 17/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab
# Print confidence interval
print(f"95% Confidence Interval for Male Purchase Mean: ({lower_bound:.2f}, {upper_bound:.2f})")

95% Confidence Interval for Male Purchase Mean: (9059.41, 9672.62)

The mean of the sampling distribution, which represents the average of all the sample means taken, is very close to the original population
mean.

keyboard_arrow_down Females:
Population Mean: 8671
Sample Mean: 8675

Males:
Population Mean: 9367
Sample Mean: 9371

This aligns with the first property of the Central Limit Theorem, which states that the sampling distribution mean approximates the population
mean.

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 18/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab
female_sample_means = [df[df['Gender']=='F'].sample(300, replace=True)['Purchase'].mean() for i in range(1000)]
male_sample_means = [df[df['Gender']=='M'].sample(300, replace=True)['Purchase'].mean() for i in range(1000)]

# Compute the 95% Confidence Intervals

female_lower, female_upper = np.percentile(female_sample_means, [2.5, 97.5])
male_lower, male_upper = np.percentile(male_sample_means, [2.5, 97.5])

# Visualization
plt.figure(figsize=(12, 6))
sns.histplot(female_sample_means, bins=30, kde=True, color="purple", label="Female Sample Means", alpha=0.6)
sns.histplot(male_sample_means, bins=30, kde=True, color="blue", label="Male Sample Means", alpha=0.6)

# Add vertical lines for confidence intervals

plt.axvline(female_lower, color='red', linestyle='dashed', label=f'Female 95% CI: {female_lower:.2f} - {female_upper:.2f}')
plt.axvline(female_upper, color='red', linestyle='dashed')
plt.axvline(male_lower, color='green', linestyle='dashed', label=f'Male 95% CI: {male_lower:.2f} - {male_upper:.2f}')
plt.axvline(male_upper, color='green', linestyle='dashed')

# Add means
plt.axvline(np.mean(female_sample_means), color='purple', linestyle='solid', label=f'Female Mean: {np.mean(female_sample_means):.2f}')
plt.axvline(np.mean(male_sample_means), color='blue', linestyle='solid', label=f'Male Mean: {np.mean(male_sample_means):.2f}')

# Labels and title

plt.xlabel("Sample Mean of Purchase")
plt.ylabel("Frequency")
plt.title("Comparison of Male vs. Female Purchase Means with 95% CI")
plt.legend()
plt.show()

# Print confidence intervals

print(f"95% Confidence Interval for Female Purchase Mean: ({female_lower:.2f}, {female_upper:.2f})")
print(f"95% Confidence Interval for Male Purchase Mean: ({male_lower:.2f}, {male_upper:.2f})")

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 19/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

95% Confidence Interval for Female Purchase Mean: (8178.71, 9206.66)

95% Confidence Interval for Male Purchase Mean: (8790.47, 9959.93)

keyboard_arrow_down Impact of Marital Status on Spending Amount

# Mapping integer values 0/1 to Unmarried/Married -
df['Marital_Status'] = df['Marital_Status'].map({0: 'Unmarried', 1: 'Married'})

# Marital Status -
x = df['Marital_Status'].value_counts().values

plt.figure(figsize=(5, 5))
plt.pie(x, center=(0, 0), radius=1.5, labels=df['Marital_Status'].unique(), autopct='%1.1f%%', pctdistance=0.5)
plt.title('Marital Status')
plt.axis('equal')
plt.show()

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 20/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

# No. of unique users in each category -

df.groupby('Marital_Status')['User_ID'].nunique()

User_ID

Marital_Status

Married 2474

Unmarried 3417

dtype: int64

# Checking different metrics based on purchase by different categories -

df.groupby('Marital_Status')['Purchase'].describe()

count mean std min 25% 50% 75% max

Marital_Status

Married 224149.0 9187.040076 4925.205232 12.0 5833.0 8042.0 12006.0 21398.0

Unmarried 323242.0 9201.581849 4948.327397 12.0 5480.0 8035.0 12028.0 21399.0

# Plotting all the observations -

sns.displot(x='Purchase', data=df, bins=25, hue='Marital_Status')

plt.axvline(x=df['Purchase'].mean(), color='r')
plt.axvline(x=df[df['Marital_Status']=='Married']['Purchase'].mean(), color='g')

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 21/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab
plt.axvline(x=df[df['Marital_Status']=='Unmarried']['Purchase'].mean(), color='b')

plt.show()

keyboard_arrow_down The distribution of Purchase appears close to normal.

Let us pick around 1,000 random samples of size 300 from the entire data set and calculate the mean of each sample.

Size = 300

Iterations = 1000

unmarried_sample_means=[df[df['Marital_Status']=='Unmarried']['Purchase'].sample(300).mean() for i in range(1000)]

married_sample_means=[df[df['Marital_Status']=='Married']['Purchase'].sample(300).mean() for i in range(1000)]

plt.figure(figsize=(10, 6))

# Plot both distributions

sns.histplot(unmarried_sample_means, bins=30, kde=True, color='blue', label="Unmarried")
sns.histplot(married_sample_means, bins=30, kde=True, color='green', label="Married")

# Add vertical lines for the population mean of each category

plt.axvline(x=df[df['Marital_Status'] == 'Unmarried']['Purchase'].mean(), color='blue', linestyle='dashed', label="Unmarried Mean")
plt.axvline(x=df[df['Marital_Status'] == 'Married']['Purchase'].mean(), color='green', linestyle='dashed', label="Married Mean")

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 22/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

plt.xlabel("Sample Mean")
plt.ylabel("Frequency")
plt.title("Comparison of Sampling Distributions (Married vs. Unmarried)")
plt.legend()
plt.show()

# Mean of the sampling distributions

mean_unmarried = np.mean(unmarried_sample_means)
mean_married = np.mean(married_sample_means)

# Standard deviation of the sampling distributions (Standard Error)

std_unmarried = np.std(unmarried_sample_means, ddof=1) # ddof=1 for sample std dev
std_married = np.std(married_sample_means, ddof=1)

print(f"Unmarried - Mean: {mean_unmarried:.2f}, Standard Deviation: {std_unmarried:.2f}")

print(f"Married - Mean: {mean_married:.2f}, Standard Deviation: {std_married:.2f}")

Unmarried - Mean: 9210.43, Standard Deviation: 276.42

Married - Mean: 9187.20, Standard Deviation: 282.31

keyboard_arrow_down Insights from the Sampling Distributions

Mean Comparison:
https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 23/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab
Unmarried Mean = 9210.43

Married Mean = 9187.20

Observation: The means are very close, indicating that both groups have similar average purchases.

Standard Deviation (Standard Error) Comparison:

Unmarried SE = 276.42

Married SE = 282.31

Observation: The standard errors are also close, suggesting that both groups have similar variability in their spending patterns.

keyboard_arrow_down To determine a range where the true mean likely falls for each group, let's calculate the 95%
Confidence Interval (CI)

# 95% Confidence Interval Calculation

ci_unmarried = stats.norm.interval(0.95, loc=mean_unmarried, scale=std_unmarried / np.sqrt(600))
ci_married = stats.norm.interval(0.95, loc=mean_married, scale=std_married / np.sqrt(600))

print(f"95% CI for Unmarried: {ci_unmarried}")

print(f"95% CI for Married: {ci_married}")

95% CI for Unmarried: (9188.308205473391, 9232.543461193274)

95% CI for Married: (9164.612804844472, 9209.791721822197)

plt.figure(figsize=(8, 6))

# Unmarried Error Bar

plt.errorbar(x='Unmarried',
y=mean_unmarried,
yerr=ci_unmarried[1] - mean_unmarried,
fmt='o', color='blue', label='Unmarried')

# Married Error Bar

plt.errorbar(x='Married',
y=mean_married,
yerr=ci_married[1] - mean_married,
fmt='o', color='green', label='Married')

plt.ylabel('Mean Purchase')
plt.title('Mean Purchase with 95% Confidence Interval')
plt.legend()
plt.show()

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 24/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

plt.figure(figsize=(8, 6))
sns.kdeplot(df[df['Marital_Status'] == 'Unmarried']['Purchase'], label='Unmarried', color='blue', fill=True)
sns.kdeplot(df[df['Marital_Status'] == 'Married']['Purchase'], label='Married', color='green', fill=True)

plt.axvline(mean_unmarried, color='blue', linestyle='dashed', label='Unmarried Mean')

plt.axvline(mean_married, color='green', linestyle='dashed', label='Married Mean')

plt.title('Purchase Distribution by Marital Status')

plt.xlabel('Purchase Amount')
plt.legend()
plt.show()

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 25/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab

keyboard_arrow_down 📌 Summary of Observations & Insights

Throughout this analysis, we explored purchase behavior across different demographic factors using statistical methods and visualizations.
Below is a structured summary of our key findings:

🔍 1. Purchase Behavior by Gender

Boxplot Analysis: Males tend to spend slightly more than females.
Distribution Analysis: Mean purchase values for males and females are close, but females spend slightly less.

📊 Sampling Distribution:
Mean of the sampled female purchases closely matches the population mean (8671 vs. 8675).
Mean of the sampled male purchases closely matches the population mean (9367 vs. 9371).
✅ This confirms the Central Limit Theorem: The sampling distribution mean approximates the population mean.

🔍 2. Purchase Behavior by Marital Status

Married vs. Unmarried Spending:

Unmarried Mean: 9210.43

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 26/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab
Married Mean: 9187.20
📌 Standard deviations are close, indicating similar purchase variability.
Distribution Analysis: The spending distribution is approximately normal for both groups.

📊 Sampling Distribution Analysis:

Sampling distribution means for both groups closely match their original population means.
✅ This further validates the Central Limit Theorem.

📊 3. Confidence Interval Analysis

95% Confidence Intervals (CI) were calculated and visualized for both Married and Unmarried groups.
Visualization Insight:

The confidence intervals overlap, suggesting no statistically significant difference in spending between married and unmarried
users.

🚀 Final Takeaways
✔ The Central Limit Theorem holds, as the sampling distribution means approximate population means.
✔ Gender-based spending is slightly higher for males, but the difference is not very large.
✔ Marital status has little impact on purchase behavior, as seen from the overlapping confidence intervals.
✔ The overall spending distribution is close to normal, making sampling-based estimations reliable.

Start coding or generate with AI.

add Code add Text

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 27/28
2/17/25, 11:35 PM Walmart_Business_Case_Study.ipynb - Colab
Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

Start coding or generate with AI.

https://colab.research.google.com/drive/1KQYVbLLPQh_jGV_DacOK1Ii8M4marGGo#scrollTo=d0KtEd2SFcOl&printMode=true 28/28