Dynamic Pricing Optimization Engine

A high-performance, data-driven microservice that calculates the mathematically optimal price for products to maximize profit. Built with FastAPI, this engine leverages Pandas and PyArrow for high-speed in-memory data caching, and SciPy to compute multivariate optimization bounds based on real-time market conditions.

The project is organized into three core layers:

1. Data Layer (data_layer.py) — Uses PyArrow to load a highly compressed Parquet database into RAM on server boot. This eliminates disk I/O bottlenecks and enables the API to filter hundreds of thousands of rows in milliseconds using Pandas Boolean indexing.

2. Optimization Engine (main.py) — Uses SciPy's minimize algorithm to locate the exact peak of a profit parabola. It balances base costs, dynamic demand multipliers, and competitor pricing to determine the absolute maximum profit margin.

3. API Routing (main.py) — Served via FastAPI and secured with Pydantic data validation.

Follow the steps below to generate the synthetic database, convert it for high-speed access, and start the API server.

git clone https://github.com/Yash49-Xe/dynamic-pricing-engine.git
cd dynamic-pricing-engine

# Create and activate a virtual environment
python -m venv venv
source venv/bin/activate        # On Windows: .\venv\Scripts\activate

# Install dependencies
pip install fastapi uvicorn pydantic scipy pandas pyarrow numpy

Do not supply an external CSV file. Run the included generator script to synthesize 100,000 realistic products, complete with base costs, competitor markups, and historical sales data.

python generate_data.py

Convert the raw CSV into a highly compressed, column-oriented Parquet file for instant RAM caching on server boot.

python convert.py

python -m uvicorn main:app --reload

Once the server is running, navigate to http://127.0.0.1:8000/docs to access the interactive Swagger UI, or send requests directly to the endpoint below.

POST /api/v1/optimize-price

The client submits only the product ID. All data extraction and computation are handled securely on the backend.

{
  "product_id": "widget_84291"
}

{
  "product_id": "widget_84291",
  "base_cost": 15.0,
  "competitor_price": 30.0,
  "optimized_price": 34.5,
  "expected_profit": 19012.5
}