Windguru Forecast DataAgent

Continuously collects wind and wave forecast data from Windguru to build an irreplaceable historical forecast archive.

Why This Matters

Weather forecast data is ephemeral - sites like Windguru overwrite predictions constantly. No one archives what the forecast was - only what actually happened. This DataAgent captures forecasts before they disappear, building a dataset that:

• Shows forecast accuracy over time
• Enables "what was predicted vs what happened" analysis
• Creates a competitive moat (no one else has this data)

Scheduled Runs

Run every 6 hours via Apify Schedules:

• Cron: 0 0,6,12,18 * * *
• This captures 4 snapshots per day of the evolving forecast

Input

Parameter	Description	Default
spots	Array of Windguru spot IDs	Top 10 wind spots
forecastHours	Hours ahead to capture (max 240)	168 (7 days)
supabaseUrl	Optional Supabase project URL	-
supabaseKey	Optional Supabase anon key	-

Priority Spots (Default)

ID	Location	Known For
43	Tarifa, Spain	Kiteboarding mecca
64	Cabarete, DR	Consistent trade winds
49045	Hood River, OR	Gorge windsurfing
3	Maui - Hookipa	World-class waves + wind
211	Cape Town, SA	Big wind, big waves
158	Fuerteventura	Year-round wind

Output Schema

{
  "capturedAt": "2026-01-26T14:30:00Z",
  "forecastDate": "2026-01-28",
  "forecastHour": 15,
  "hoursAhead": 48,
  "spotId": "43",
  "spotName": "Tarifa, Spain",
  "windSpeed": 25,
  "windGusts": 32,
  "windDirection": 110,
  "waveHeight": 1.2,
  "wavePeriod": 8,
  "waveDirection": 270,
  "temperature": 18,
  "cloudCover": 20,
  "precipitation": 0,
  "forecastModel": "GFS",
  "source": "windguru"
}

Local Testing

cd /home/peteylinux/Projects/Apify/dataagents/windguru-forecast-agent
npm install
mkdir -p storage/key_value_stores/default
echo '{"spots": ["43"]}' > storage/key_value_stores/default/INPUT.json
npm start

Deploy to Apify

cd /home/peteylinux/Projects/Apify/dataagents/windguru-forecast-agent
apify push

Then configure a schedule in Apify Console.

Supabase Schema

If using Supabase, create this table:

CREATE TABLE forecasts (
  id SERIAL PRIMARY KEY,
  source VARCHAR(50) NOT NULL,
  captured_at TIMESTAMPTZ NOT NULL,
  forecast_date DATE,
  hours_ahead INTEGER,
  location_id VARCHAR(100),
  location_name VARCHAR(200),
  wind_speed DECIMAL(5,1),
  wind_gusts DECIMAL(5,1),
  wind_direction INTEGER,
  wave_height DECIMAL(4,1),
  wave_period DECIMAL(4,1),
  wave_direction INTEGER,
  temperature INTEGER,
  cloud_cover INTEGER,
  precipitation DECIMAL(5,1),
  forecast_model VARCHAR(50),
  raw_json JSONB,
  created_at TIMESTAMPTZ DEFAULT NOW()
);

CREATE INDEX idx_forecasts_lookup ON forecasts(source, location_id, forecast_date);
CREATE INDEX idx_forecasts_captured ON forecasts(source, captured_at);