This project implements a reproducible ETL pipeline and EDA workflow for analyzing e-commerce order data.

├── data/
│   ├── raw/            # Input CSVs (orders.csv, users.csv)
│   └── processed/      # Output Parquet files + run metadata
├── notebooks/
│   └── eda.ipynb       # Exploratory Data Analysis
├── reports/
│   ├── figures/        # Exported plots
│   └── summary.md      # Executive summary of findings
├── scripts/
│   └── run_etl.py      # Main entrypoint for the pipeline
└── src/
    └── bootcamp_data/  # Python package with ETL logic

1. Setup Environment: Ensure you have the required dependencies installed (pandas, plotly, pyarrow).

2. Run the ETL Pipeline: This script reads raw data, cleans it, joins it with users, and saves processed Parquet files.

   python scripts/run_etl.py

3. Check Outputs: Processed files will appear in data/processed/:

   • analytics_table.parquet: The main joined table for analysis.
   • _run_meta.json: Metadata about the last run (row counts, paths).

4. Run Analysis: Open notebooks/eda.ipynb to view the analysis and visualizations.

The pipeline generates the following artifacts in data/processed/:

• analytics_table.parquet: Enriched orders with user details and time features.
• orders_clean.parquet: Cleaned orders table.
• users.parquet: Cleaned users table.