Weather Analysis

# Weather Analysis

Correlate subjective weather experience with objective meteorological data. Interview first, then fetch data, then explain what happened.

## Arguments

`<location> <date>` — both optional.

- `/weather-analysis` — Iowa City, today
- `/weather-analysis NYC yesterday` — New York, yesterday
- `/weather-analysis 2026-05-05` — Iowa City, May 5

Default location: **Iowa City**
Default date: **today**

## Workflow

### Step 1: Interview

Ask the user what they experienced. Keep it open-ended — one or two questions max:

- "What did the weather feel like today? What stood out — anything uncomfortable, surprising, or that changed suddenly?"
- If they mention a specific time window, ask about that: "What happened around [time] — did it shift suddenly?"

Wait for their response. Do NOT proceed until you have their subjective account.

If the user already provided their experience in the same message as the skill invocation (or in a recent voice note), skip the interview — you already have the context.

### Step 2: Run Weather Report

Invoke the `/weather-report` skill for the resolved location and date. This fetches the full meteorological data: daily summary, hourly breakdown, and weather narrative.

### Step 3: Correlate and Explain

Map the user's subjective experience to the objective data. This is the core of the skill. For each thing they described:

1. **Identify the time window** they're talking about
2. **Pull the specific hourly data** for that window
3. **Explain the physics** of why it felt the way it did

Examples of correlations:
- "You said it got bad around 1 PM — that's exactly when cloud cover dropped from 95% to 12% and UV jumped to 6.8. Direct solar radiation adds 10-15°F to perceived temperature."
- "The step-function sun/shadow you felt matches the cloud data: 12% at 1 PM, 5% at 2 PM, 99% at 3 PM, 10% at 4 PM."
- "The residual heat after sunset — wind collapsed from 12 mph to 2 mph while humidity climbed from 35% to 54%. The ground was radiating stored thermal energy with no convective cooling."

### Step 4: Summary

End with 2-3 sentences summarizing the key insight — what was the dominant factor that made the day feel the way it did? Was it the sun, the wind, the humidity, a transition, a combination?

Be direct and specific. "It was hot" is not an insight. "The erratic cloud cover created a 15°F perceived temperature oscillation every hour while declining wind removed the only cooling mechanism by evening" is.

## Rules

1. **Always interview first.** The user's subjective experience is the anchor — the data explains it, not the other way around.
2. **Be specific about time.** Don't say "in the afternoon." Say "at 2 PM."
3. **Explain the physics.** Why did it feel that way? Solar radiation, convective cooling, thermal mass, humidity trapping — name the mechanism.
4. **No lifestyle advice.** This skill explains weather, it doesn't tell people what to do about it. That's what `/checkweather-vivek` is for.