I've built SalesPro AI, an interactive, full-stack Next.js web application designed to help businesses instantly forecast sales performance based on their marketing spend across TV, Radio, and Newspaper advertising budgets. The core of the app runs a custom-trained Polynomial Regression Model (achieving 93.55% accuracy / R² score) directly in the browser. By leveraging TypeScript to run the ML math client-side, we get instantaneous forecasts with zero API latency or backend overhead. 🛠️ Key Features: ML-Powered Forecaster: Instant sales projections and customized channel allocation insights based on real historical marketing data. Interactive Analytics Dashboard: Real-time stats tracking total, average, and highest predictions. Persistent Session & History: Secure local signup, login, demo access, and custom settings (including dynamic profile picture uploads!). Data Portability: Full history logs with sorting, filtering, and a one-click Export to CSV feature for further spreadsheet analysis. Modern UX/UI: Responsive dark-mode dashboard with smooth animations and clean micro-interactions. 💻 The Tech Stack: Frontend: Next.js 14 (App Router) & React Styling: Tailwind CSS & Modern Glassmorphism Logic/Type Safety: TypeScript Algorithms: Polynomial Regression (Degree 2) Storage: LocalStorage-backed authentication & prediction database 👉 Live Demo: [https://lnkd.in/eUM_zKkQ]