INDUSTRIAL-PC-SQL-FEATURE-ML

Order-to-delivery demand analysis project using SQL-based feature engineering and baseline modeling.

Project Overview

This repository contains an end-to-end workflow:

1. Generate synthetic/structured data into a SQLite database
2. Build features via SQL aggregations
3. Validate a baseline model and evaluation metrics

Repository Structure

• notebooks/
  • 01_eda_order_to_delivery.ipynb — EDA & problem framing
  • 02_feature_engineering.ipynb — feature design & validation
  • 03_model_validation.ipynb — baseline model & evaluation
• src/
  • generate_customers.py — generate customer-related tables
  • generate_products_orders.py — generate products/orders tables
  • step1_sql_features.py — SQL feature engineering pipeline
  • check_db.py — database sanity checks
• data/
  • raw/ — SQLite DB (excluded from git)
  • processed/ — feature artifacts (excluded from git)
• outputs/ — figures/results (optional)

Note: Data files (*.db, *.parquet) are excluded from version control via .gitignore.

How to Run (Quick Start)

1) Create environment & install dependencies

pip install -r requirements.txt

2) Generate database (raw data)

python src/generate_customers.py
python src/generate_products_orders.py

3) Build features

python src/step1_sql_features.py

4) Open notebooks

jupyter notebook

Outputs

Figures: outputs/figures/ (if enabled in notebooks)

Tech Stack

• Python (pandas, numpy)
• SQLite
• Jupyter Notebook
• scikit-learn (baseline modeling)

License

For portfolio / demonstration use.