E-Commerce Data Analysis

A comprehensive analysis and interactive dashboard for e-commerce data, featuring both Jupyter notebook exploration and a professional Streamlit dashboard.

Overview

This project provides tools for analyzing e-commerce performance metrics including revenue trends, customer behavior, product performance, and geographic distribution of sales. The analysis examines data from 2022-2023 with the ability to filter by custom date ranges.

Contents

• EDA_Refactored.ipynb - Comprehensive Jupyter notebook with exploratory data analysis
• dashboard.py - Interactive Streamlit dashboard with real-time filtering
• data_loader.py - Utility functions for loading and preparing datasets
• business_metrics.py - Functions for calculating key business metrics
• ecommerce_data/ - Raw CSV datasets directory

Features

Dashboard Highlights

The Streamlit dashboard provides a professional, interactive interface with:

Header Section

• Application title
• Global date range filter (applies to all visualizations)

KPI Row (4 Cards)

• Total Revenue - Total sales with year-over-year growth trend
• Monthly Growth - Month-over-month revenue change percentage
• Average Order Value (AOV) - Mean transaction value with trend indicator
• Total Orders - Order volume with trend comparison

Each KPI card displays:

• Large, readable metric value
• Trend indicator (↑ for growth, ↓ for decline)
• Percentage change with color coding (green for positive, red for negative)

Analytics Charts (2x2 Grid)

1. Revenue Trend Line Chart

   • Solid line for current period
   • Dashed line for previous period (comparison)
   • Grid lines for easier reading
   • Y-axis formatted as $300K, $2M, etc.

2. Top 10 Categories Bar Chart

   • Horizontal bar chart sorted in descending order
   • Blue gradient coloring based on revenue amount
   • Values formatted as $300K, $2M

3. Revenue by State (US Choropleth Map)

   • Color-coded by revenue amount
   • Blue gradient color scale
   • Interactive hover details
   • Albers USA projection for accurate representation

4. Customer Satisfaction vs Delivery Time

   • Bar chart showing average review scores
   • Delivery time buckets (1-3 days, 4-7 days, etc.)
   • Demonstrates correlation between faster delivery and higher satisfaction

Bottom Row (2 Cards)

1. Average Delivery Time

   • Large display of average days to delivery
   • Trend indicator showing improvement/decline vs previous period
   • Color-coded trend arrow

2. Review Score

   • Large average review score (out of 5.0)
   • Star rating display
   • Subtitle: "Average Review Score"

Data Dictionary

Key Tables

orders_dataset.csv

• order_id - Unique identifier for each order
• customer_id - Identifier linking to the customer
• order_status - Current status (canceled, delivered, pending, processing, shipped, returned)
• order_purchase_timestamp - Date and time order was placed
• order_delivered_customer_date - Date order was delivered

order_items_dataset.csv

• order_id - Links to orders table
• product_id - Links to products table
• price - Price of individual item
• freight_value - Shipping cost for item

products_dataset.csv

• product_id - Unique product identifier
• product_category_name - Product category
• product_description_length - Description length

customers_dataset.csv

• customer_id - Unique customer identifier
• customer_state - State location (US states)
• customer_city - City location

order_reviews_dataset.csv

• review_id - Unique review identifier
• order_id - Links to orders table
• review_score - Rating from 1-5
• review_creation_date - When review was posted

Installation

1. Clone or download the repository

2. Install dependencies

   pip install -r requirements.txt

3. Ensure data files are in place

   • Place CSV files in the ecommerce_data/ directory

Usage

Running the Streamlit Dashboard

streamlit run dashboard.py

The dashboard will open in your browser at http://localhost:8501. Use the date range filter in the header to dynamically update all visualizations.

Running the Jupyter Notebook

jupyter notebook EDA_Refactored.ipynb

The notebook provides detailed exploratory analysis with configuration parameters at the top for adjusting the analysis period.

Key Metrics

The dashboard calculates and displays:

• Total Revenue - Sum of all order values
• Revenue Growth - Year-over-year percentage change
• Average Order Value (AOV) - Mean revenue per order
• Total Orders - Count of delivered orders
• Average Delivery Speed - Days from order to delivery
• Average Review Score - Customer satisfaction rating (1-5 scale)
• Revenue by Category - Sales distribution across product categories
• Revenue by State - Geographic sales distribution
• Delivery Time vs Satisfaction - Correlation analysis

Technical Details

Technologies Used

• Streamlit - Interactive dashboard framework
• Plotly - Interactive charting library
• Pandas - Data manipulation and analysis
• NumPy - Numerical computing
• Matplotlib - Static visualization (EDA notebook)

Styling

• Professional card-based layout
• Consistent color scheme (blue primary, red/green for trends)
• Responsive design that adapts to different screen sizes
• Custom HTML/CSS for KPI cards
• Plotly's built-in interactivity for charts

Analysis Insights

Based on 2022-2023 data analysis:

1. Revenue Performance - Slight decline of 2.46% year-over-year
2. Order Volume - 2.40% decrease in order count
3. AOV Stability - Average order value remained stable at ~$725
4. Customer Satisfaction - Consistent 4.1/5.0 average review score
5. Delivery Performance - Maintained ~8 days average delivery
6. Geographic Distribution - Sales across all US states with California leading

Customization

Adjusting Date Range

Use the date filters in the dashboard header to analyze any period within your dataset.

Modifying Chart Colors

Edit the color definitions in dashboard.py:

• COLOR_PRIMARY - Primary chart color
• Plotly colorscales for gradients

Adding New Metrics

Use the calculate_* functions in business_metrics.py to add new KPIs or charts.

Troubleshooting

Issue: Data not loading

• Ensure CSV files are in the ecommerce_data/ directory
• Check that filenames match exactly in data_loader.py

Issue: Charts not displaying

• Verify Plotly is installed: pip install --upgrade plotly
• Clear Streamlit cache: streamlit cache clear

Issue: Date filter not affecting charts

• Check that the date range is within your dataset's date boundaries
• Verify data preparation filters in prepare_data() function

Future Enhancements

• Customer segmentation analysis
• Predictive revenue forecasting
• Product performance drill-down
• Customer lifetime value calculations
• Inventory optimization insights
• Marketing attribution analysis

License

This project is provided as-is for educational and analytical purposes.

Support

For questions or issues, refer to the inline documentation in:

• business_metrics.py - Metric calculation functions
• data_loader.py - Data loading and preparation
• dashboard.py - Dashboard structure and components