Clear Skies: Building a Backyard Weather App

The problem was not missing data. I already had Astrospheric for astronomy-specific signals, Weather.gov for authoritative local forecasts, and Carrot Weather for quick checks. The friction was the workflow: too many tabs, too many paid upsells, and too much context switching for one decision.

Start with the decision

The app had one job: tell me whether the rig should go outside, stay outside through the morning, or stay packed. Once that decision was clear, the data and UI could follow.

Pick the signals that matter

The first research pass split the forecast into practical questions: rain in the next 24 hours, cloud cover when dark, seeing guidance, moon interference, and a multi-night outlook.

• NWS precipitation and grid data
• Open-Meteo cloud layers and model blending
• Optional 7Timer seeing guidance
• Moon calculations for observing context

Keep the backend opinionated

Python gathers NWS, Open-Meteo, and optional 7Timer data, blends the hourly forecast, applies scope rules, then returns browser-ready JSON. That lets the frontend stay focused on presentation rather than weather math.

Make the scan path short

The UI uses a collapsible headline, live cloud map, 72-hour chart, and next-seven-nights strip so the useful answer arrives first and the detail stays one click away.

What agentic development changed

I used AI agents to move through research, spikes, refactors, deployment plumbing, UI polish, and tests without losing product judgment. The human constraint stayed stable: do not build a weather portal; build my backyard observing answer.

What worked and what still needs work

The strongest choice was keeping the app opinionated. The remaining challenge is confidence: making model agreement, freshness, and uncertainty obvious without turning the page back into a dense weather dashboard.

Further reading

If you want the presentation version of the story, the case study deck walks through the same project as a slide sequence.