[Keld] Add predictive correlation feature for environmental factors

I'm an electrical engineering student (heading into biomedical engineering for grad school) who started tracking correlations between Apple Watch data (HRV, sleep stages, respiratory rate, SpO2, activity) and environmental factors (barometric pressure, humidity, air quality, temperature). What began as a class-adjacent obsession turned into a full app. The interesting finding: some correlations aren't just reactive — they're predictive. My deep sleep architecture shifts *before* barometric pressure drops, not after. My HRV dips hours before air quality degrades. The body seems to telegraph what the atmosphere is about to do. The app is called Keld. It reads Apple Health + WeatherKit and runs a correlation engine entirely on-device. The main visualization is what we call the Elemental Bond Map — it draws every statistically significant connection between your biology and your environment. Gold curves = signals that move together, blue = opposition, thickness = strength. The part that keeps me hooked: everyone's map looks completely different. The patterns are genuinely personal — shaped by where you live, how you sleep, what you're sensitive to. Everything runs on-device. Your raw Apple Health readings never leave your phone. If you opt into the community feature, only anonymized pattern summaries are shared — never individual readings. We're running a first wave beta with 100 TestFlight spots open now. A second wave of 1,000 will open when we're ready. Looking specifically for people who already track seriously and would notice things I wouldn't. Requires iPhone + Apple Watch (or any wearable connected to Apple Health). TestFlight: [https://testflight.apple.com/join/fzb8wDJ6](https://testflight.apple.com/join/fzb8wDJ6) Happy to answer questions about the correlation methodology or anything else.