Using Python for Analyzing Purchase Behavior on Black Friday – Walmart Inc.

Overview

This project investigates customer purchasing behavior at Walmart Inc., focusing on gender, age, and marital status. By analyzing a dataset of over 550,000 transactions, we explore spending patterns and trends to provide actionable insights for Walmart's strategic decision-making.

Key Objectives

1. Understand demographic-based purchasing behavior: Explore variations in spending across gender, age groups, and marital status.
2. Analyze key trends: Identify high-spending demographics and popular products.
3. Provide actionable insights: Assist Walmart in improving marketing strategies, inventory planning, and customer experience.

---

Dataset Information

• Size: 550,068 rows, 10 attributes.
• Key Attributes:
  • Customer Demographics: Gender, Age, Marital Status, City Category.
  • Purchase Data: Purchase Amount, Product Codes.

---

Methodology

1. Preprocessing

• Data Cleaning: Removal of null values and duplicates.
• Outlier Detection: Utilized Interquartile Range (IQR) for identifying and addressing outliers.
• Feature Encoding: Encoded categorical variables for analysis.
• Exploratory Data Analysis (EDA): Basic statistics and distribution checks.

2. Univariate Analysis

• Techniques: Bar plots, pie charts, count plots.
• Focus:
  • Gender distribution.
  • Age demographics.
  • Marital status proportions.
  • Popular products by gender.

3. Bivariate Analysis

• Techniques: Box plots, line graphs.
• Focus:
  • Spending trends based on gender and age.
  • Impact of marital status on spending patterns.
• Correlation Analysis: Assessed relationships between categorical and numerical attributes.

4. Statistical Analysis - Central Limit Theorem (CLT)

• Bootstrap Sampling: Enhanced reliability by constructing confidence intervals.
• Confidence Intervals: Estimated population mean for spending patterns.

---

Key Findings

1. Gender-Based Insights

• Men constitute 75% of the customer base and spend more on average than women.
• Confidence Interval (95%):
  • Male Spending: $9,407 to $9,469.
  • Female Spending: $8,705 to $8,764.

2. Age-Based Insights

• Highest spending group: 50-55 years.
• Lowest spending group: 0-17 years.

3. Marital Status-Based Insights

• Unmarried customers spend slightly more than married customers.
• Spending gap is minimal, indicating marriage is not a significant factor.

4. Popular Products

• Certain product categories show distinct preferences by gender and age group, influencing inventory planning.

---

Visualizations

Univariate Analysis

• Bar and Pie Charts: Gender, age, and marital status distributions.

Bivariate Analysis

• Box Plots: Spending trends across demographics.
• Line Graphs: Average spending trends by age.

Statistical Analysis

• Confidence intervals to validate findings and highlight spending disparities.

---

Tools and Technologies

• Programming Language: Python
• Libraries: Pandas, NumPy, Matplotlib, Seaborn.
• Statistical Techniques: Central Limit Theorem, Bootstrap Resampling.

---

Conclusion

This project reveals significant insights into Walmart's customer purchasing patterns:

• Gender: Men spend more consistently than women.
• Age: The 50+ age group spends the most.
• Marital Status: Differences between married and unmarried customers are minimal.

These findings empower Walmart to develop data-driven strategies to improve marketing campaigns, personalize customer experiences, and optimize inventory management.

---

References

1. Tengli, A., Srinivasan, S. H. (2022). Gender-Based Purchase Behavior Study on Natural Cosmetics.
2. Pakasi, A., Tumiwa, J. (2016). Comparison Analysis of Male and Female Consumer Purchase Behavior.
3. Ronaghi, M., Danae, H., & Haghtalab, H. (2013). Effects of Gender on Consumer Behavior.
4. Abir Smiti (2020). Outlier Detection Methods in Data Analysis.