Dashboard

🚀 SMSU Hacks – AI IT Cost Management System

🌟 Inspiration & Purpose

This project was inspired by real challenges in IT budgeting, where teams struggle to track multi-year costs, adjust for inflation, and onboard new members. We built a Cost Management System that not only tracks planned vs. actual expenses but also uses an AI chatbot to explain data and guide users.

🧠 What We Learned

Full-stack development using React, Node.js, and SQLite
Integrating a local LLM (Ollama/LM Studio) for privacy-focused AI
Designing clean database schemas and REST APIs
Building interactive dashboards with Recharts
Creating a chatbot that explains real backend data

🏗️ How We Built It

Frontend: React 19 + TypeScript + Vite + Tailwind CSS
Backend: Express.js + SQLite
AI Layer: Local LLM connected via backend proxy
Features:
- Planned vs Actual cost tracking
- 5-year forecast with inflation
- Dashboard analytics via /api/analytics
- AI chatbot for explanation, training, and Q&A

Forecast formula used:

[ \text{Forecast} = \text{Base Cost} \times (1 + r)^n ]

⚡ Challenges We Faced

AI + Data Integration: Ensuring the chatbot uses real backend data without hallucination
Local LLM Limitations: Optimizing prompts and context size
Consistent Analytics: Unifying dashboard data using a single 5-year forecast source
CORS Issues: Solved by proxying AI requests through the backend