Dynamic Price Optimization Recommendation Engine

Sprint 0 – Planning & Setup Documentation

---

1. Project Overview

This project aims to build a Dynamic Price Optimization Engine that recommends optimal markdowns for low-performing retail products.
The system analyzes sales velocity, inventory availability, competitor pricing, and demand elasticity to produce data-backed price recommendations.

The final output includes:

• Optimized markdown percentages
• New recommended prices
• Impact estimation (before/after KPIs)
• A Streamlit dashboard for visualization

---

2. Problem Statement Understanding

Retailers face multiple challenges:

• Products selling slower than expected
• Excess inventory occupying working capital
• Competitors offering lower prices
• Lack of timely markdown strategies

The problem requires building an automated system that:

1. Ingests product, sales, inventory, and competitor datasets
2. Engineers features like velocity, stock health, competitiveness
3. Applies pricing logic
4. Outputs data-backed recommendations

---

3. Finalized Solution Approach

Core Components

1. Data Ingestion Layer
   Load Kaggle datasets for products, sales, inventory, and competitor prices.
   Standardize columns and unify them into a single master dataset.

2. Feature Engineering Layer

   • Weekly sales velocity
   • Inventory health (stock-to-demand ratio)
   • Competitiveness score
   • Elasticity indicator

3. Price Optimization Engine
   Rule-based strategy:

   • Identify low-performing SKUs
   • Apply markdown percentages based on demand, stock, and competitor gap
   • Estimate improvements

4. Recommendation Generator
   Outputs:

   • Markdown %
   • New price
   • Review window (7–14 days)

5. Dashboard Layer
   A Streamlit dashboard for:

   • Visual exploration
   • SKU-level recommendations
   • Downloadable CSV output

---

4. Dataset Selection and Mapping

Requirement	Kaggle Dataset
Product catalog	Retail Product Dataset
Sales line items	Sample Sales Dataset
Inventory snapshot	Product Inventory Dataset 2025
Competitor pricing	Competitor Pricing Dataset

Additional dataset documentation will be placed inside /data/.

---

5. High-Level Architecture

Data Sources (Kaggle CSVs)
        |
        v
Data Ingestion & Cleaning
        |
        v
Feature Engineering Layer
        |
        v
Price Optimization Engine
        |
        v
Recommendation & Impact Module
        |
        v
Streamlit Dashboard / CSV Output

A visual architecture diagram will be placed under /docs/.

---

6. Project Roadmap

Backend

• Dataset ingestion scripts
• Data merging and cleaning
• Feature engineering
• Pricing engine implementation
• KPI computation
• Unit testing

Frontend (Dashboard)

• Streamlit setup
• UI layout for filters and tables
• Visualizations (velocity vs stock, competitor gaps)
• CSV export

Deployment

• Environment setup
• Requirements.txt
• Optional Streamlit Cloud deployment

---

7. Tech Stack

• Python
• pandas
• numpy
• matplotlib / seaborn
• scikit-learn (optional)
• Streamlit

---

8. Repository Structure

/
│── backend/
│     ├── ingestion/
│     ├── feature_engineering/
│     ├── pricing_engine/
│     ├── utils/
│
│── dashboard/
│     ├── app.py
│
│── data/
│     ├── raw/
│     ├── processed/
│
│── docs/
│     ├── architecture.png
│     ├── sprint0_readme.pdf
│
│── notebooks/
│     ├── main_pipeline.ipynb
│
│── README.md
│── requirements.txt

---