E-Commerce Customer Spending Analysis

Overview

This project analyzes customer spending patterns in an e-commerce platform using R and statistical modeling techniques. The analysis focuses on identifying key factors that influence annual customer spending.

Dataset

The dataset contains the following variables:

• Time on Website (minutes)
• Time on App (minutes)
• Average Session Length (minutes)
• Length of Membership (years)
• Yearly Amount Spent (USD)

Analysis Approach

Exploratory Data Analysis

• Visualized relationships between engagement metrics and spending
• Examined distributions of key variables
• Analyzed pairwise correlations between all numerical variables

Statistical Modeling

Primary Model: Linear Regression

• Simple linear regression predicting Yearly Amount Spent from Length of Membership
• Multiple linear regression incorporating all engagement metrics

Model Validation

• 80/20 train-test split
• Calculated performance metrics:
  • Root Mean Squared Error (RMSE)
  • Mean Absolute Percentage Error (MAPE)
  • R-squared (R2)

Key Findings

1. Membership duration is the strongest predictor of customer spending (β = $64.22 per year)
2. Time spent on mobile app shows stronger correlation with spending than website time
3. The multiple regression model explains 98% of spending variation (R2 = 0.98)
4. Model achieves $9.97 RMSE on test data

Usage

To reproduce this analysis:

1. Install required packages:

install.packages(c("ggplot2", "knitr", "scales"))

2. Run the Quarto document:

quarto render analysis.qmd

Files

• analysis.qmd: Main analysis document
• data/ecomdata.csv: Dataset
• _quarto.yml: Configuration file

Dependencies

• R (v4.0+)
• Quarto (v1.2+)
• R Packages:
  • ggplot2
  • knitr
  • scales