SkySense

SkySense combines live weather data, a small ML model, and an LLM into one outdoor-planning dashboard. The FastAPI backend aggregates current conditions and forecasts from multiple providers, runs a scikit-learn model that scores outdoor comfort from 0 to 100 with confidence intervals, and generates a daily brief with Groq-hosted Llama 3.1 that is grounded exclusively in the live data rather than model priors. A conversational assistant, Nimbus, answers from the same current conditions and ML outputs. The React 19 and TypeScript frontend renders animated hourly and 7-day views, city search over geocoding, and 7-day analytics, with in-memory TTL caching and a backup Groq key for resilience.