📈 Sales Forecasting System

AI-powered demand prediction & business intelligence dashboard
Built with Python, Flask, scikit-learn, and Chart.js

---

🔍 About

A modular Sales/Demand Forecasting System that uses historical business data (Superstore dataset) to predict future sales and present results through an interactive, business-friendly dashboard.

The system helps businesses:

• 📦 Plan inventory based on predicted demand
• 💰 Manage cash flow with revenue projections
• 👥 Prepare staffing for peak seasons
• 📉 Avoid overstocking or losses through trend analysis

---

🏗️ Architecture

The project follows a modular architecture with separate frontend and backend directories:

FUTURE_ML_01/
├── backend/                    # Python ML Pipeline + REST API
│   ├── app.py                  # Flask API server (4 endpoints)
│   ├── data_preprocessing.py   # Data cleaning & feature engineering
│   ├── model.py                # Linear Regression model & forecasting
│   ├── requirements.txt        # Python dependencies
│   └── model_artifacts/        # Serialized model (auto-generated)
│
├── frontend/                   # Interactive Dashboard (SPA)
│   ├── index.html              # Multi-page dashboard layout
│   ├── styles.css              # Premium dark-mode stylesheet
│   └── app.js                  # Chart.js visualizations & page router
│
├── Sample - Superstore.csv     # Dataset (9,994 transactions)
├── vercel.json                 # Vercel deployment config
└── README.md

---

✨ Features

📊 Dashboard (5 Pages)

Page	Description
Overview	6 KPI cards (Sales, Profit, Orders, AOV, Margin, Items) + YoY Growth chart
Forecast	Historical vs. 12-month predicted sales line chart + detail table
Breakdown	Category, Region, Segment doughnuts + Top Sub-Categories bar chart
Seasonality	Monthly patterns with holiday highlighting + yearly comparison lines
Model	Linear Regression metrics (MAE, RMSE, MAPE, R²) + Actual vs Predicted plot

🤖 ML Pipeline

• Data Preprocessing: Date parsing, aggregation, duplicate removal
• Feature Engineering: Lag features, rolling means, cyclical month encoding, holiday flags
• Model: Linear Regression with time-ordered 80/20 train/test split
• Forecasting: Iterative 12-month ahead predictions with lag propagation

🎨 Design

• Premium dark mode with glassmorphism effects
• Micro-animations (KPI counter animation, page transitions, hover effects)
• Fully responsive (desktop, tablet, mobile)
• Google Fonts (Inter) for modern typography

---

🚀 Quick Start

Prerequisites

• Python 3.8+
• pip

Installation

# Clone the repository
git clone https://github.com/Ayon-coder/FUTURE_ML_01.git
cd FUTURE_ML_01

# Install backend dependencies
cd backend
pip install -r requirements.txt

Running Locally

# Terminal 1 — Start the backend API
cd backend
python app.py
# → API running at http://127.0.0.1:5000

# Terminal 2 — Serve the frontend
cd frontend
python -m http.server 8080
# → Dashboard at http://127.0.0.1:8080

Open http://127.0.0.1:8080 in your browser.

---

🔌 API Endpoints

Method	Endpoint	Description
GET	/api/health	Health check
GET	/api/forecast	Historical + 12-month predicted sales
GET	/api/metrics	Model evaluation metrics (MAE, RMSE, MAPE, R²)
GET	/api/insights	KPIs, category/region/segment breakdowns, seasonality

---

📁 Dataset

Sample - Superstore.csv — a widely used retail dataset containing:

• 9,994 transactions across 4 years (2014–2017)
• 21 columns: Order Date, Sales, Profit, Quantity, Category, Region, Segment, etc.
• 3 Categories: Furniture, Office Supplies, Technology
• 4 Regions: Central, East, South, West
• 3 Segments: Consumer, Corporate, Home Office

---

🛠️ Tech Stack

Layer	Technology
Backend	Python, Flask, Flask-CORS
ML	scikit-learn (Linear Regression)
Data	pandas, NumPy
Frontend	HTML5, CSS3, JavaScript (ES6+)
Charts	Chart.js 4.4
Fonts	Google Fonts (Inter)

---

📄 License

This project is licensed under the MIT License.