AsteroidSafe

Web dashboard that ingests NASA NeoWs / JPL SBDB data and classifies Near-Earth Objects as Potentially Hazardous (PHA) using a pretrained tabular foundation model (TabPFN), deployed via ONNX Runtime in .NET.

---

Tech stack

• Frontend: React + TypeScript (Vite)
• Backend: ASP.NET Core Web API
• ML: Python (TabPFN + baselines)
• Deployment: ONNX Runtime (.NET)

---

Architecture

• frontend/ — dashboard UI
• backend/ — REST API + ONNX inference
• training/ — data ingestion + training + export to ONNX
• docs/ — documentation and reports

flowchart LR
  A["NASA NeoWs + JPL SBDB"] --> B["training (Python)<br/>ingestion + dataset + TabPFN"]
  B --> C["ONNX model artifact"]
  C --> D["backend (ASP.NET Core)<br/>ONNX Runtime inference API"]
  D --> E["frontend (React + TS)<br/>dashboard"]

Loading

---

Project management

• 📌 Milestones
• ✅ Issues
• 🧭 Project board

Roadmap (Milestones)

• M0 — Setup & Standards
• M1 — Data Ingestion
• M2 — Dataset & Feature Engineering
• M3 — Model Training
• M4 — ONNX + .NET Inference
• M5 — Web Dashboard
• M6 — Docs & Release