A comprehensive business analysis of 2 years of e-commerce transaction data from a Brazilian marketplace, analyzing $15.8M in revenue across 99,441 orders.
An e-commerce platform needs to understand:
- Revenue trends and growth patterns
- Customer behavior and retention issues
- Product category performance
- Geographic market distribution
- Delivery operations efficiency
- Payment method preferences
- Total Revenue: $15.8 Million (2016-2018)
- Total Orders: 99,441 orders
- Average Order Value: $159.33
- Peak Month: November 2017 - $1.18M (Black Friday effect)
- 0% Repeat Purchase Rate - No customer loyalty, massive revenue loss
- Geographic Concentration - 62.5% revenue from only 3 states
- Late Deliveries - Some states have 16-23% late delivery rates
- Missing Data - 2,378 products without category classification
- Implement customer retention program: +$1.5M potential annual revenue
- Expand to underserved regions: +15% revenue growth
- Optimize delivery logistics: +15% customer satisfaction
- Leverage installment payments: +8% average order value
- Python 3.11: Data processing and analysis
- Pandas & NumPy: Data manipulation
- Matplotlib & Seaborn: Data visualization
- Jupyter Notebook: Interactive analysis
- Git: Version control
ecommerce-sales-analysis/
├── data/
│ ├── raw/ # Original CSV files
│ └── processed/ # Cleaned master dataset
├── notebooks/
│ ├── 01_data_exploration.ipynb # Data loading & preparation
│ └── 02_business_analysis.ipynb # Business insights & visualizations
├── reports/ # Generated charts and reports
├── src/ # Python scripts (future)
└── README.md
- Clear growth pattern from 2017-2018
- Seasonal peaks in November (Black Friday)
- Stable $1M+ monthly revenue in mature period
- Beauty & Health leads with $1.44M (9.2%)
- Top 5 categories: 40% of total revenue
- Opportunity to expand successful categories
- São Paulo (SP): 37.4% of revenue
- Top 3 states (SP, RJ, MG): 62.5% of revenue
- Significant growth potential in other regions
- 92.6% on-time delivery (excellent)
- Average delivery: 12 days
- Regional improvement needed (AL, MA, SE)
Problem: 0% repeat customers Solution:
- Launch loyalty program
- Email remarketing campaigns
- Second purchase discounts Expected Impact: +$1.5M annual revenue
Problem: Revenue concentration in 3 states Solution:
- Improve logistics in underserved regions
- Local seller partnerships
- Regional marketing campaigns Expected Impact: +15% revenue growth
Problem: High late delivery rates in northeast Solution:
- Regional carrier partnerships
- Realistic delivery estimates
- Seller network expansion Expected Impact: <5% late deliveries, improved ratings
Source: Brazilian E-Commerce Public Dataset by Olist Period: September 2016 - October 2018 Records: 113,425 order items Size: 124.97 MB (processed)
- Clone the repository
- Install requirements:
pip install -r requirements.txt - Open Jupyter notebooks in
notebooks/folder - Run cells sequentially
Priyanka Malavade
- BCA Graduate | GUVI Certified Data Scientist (2025)
- Email: priyasmalavade@gmail.com
- LinkedIn: linkedin.com/in/priyanka-malavade-b34677298
- GitHub: github.com/priyanka7411
This project is for portfolio and educational purposes.
The raw dataset is not included in this repository due to size constraints (58MB+).
To reproduce this analysis:
- Download the dataset from Kaggle - Brazilian E-Commerce Dataset
- Extract the CSV files to
data/raw/archive (29)/ - Run the notebooks in order
The processed master dataset (data/processed/master_sales_data.csv) is included for reference.