Parsing Weather Data

Extract structured weather information from raw web content, HTML, or unformatted text.

Python
# Extract current conditions
location = "Panama City, Panamá"
current_temp = "29°C"
feels_like = "31°C"
conditions = "Cloudy"
wind = "E 26 km/h"

# Format as structured data
weather_data = {
    "location": location,
    "current": {
        "temperature": current_temp,
        "feels_like": feels_like,
        "conditions": conditions,
        "wind": wind
    }
}

Progress:

• Identify location and timestamp
• Extract current conditions (temp, feels-like, conditions, wind)
• Parse hourly forecast data
• Extract daily forecast (highs/lows, precipitation chance)
• Capture additional metrics (air quality, sunrise/sunset)
• Format as structured output

Example 1: Current Conditions Input: "Panama City, Panamá 29°C RealFeel® 31° Cloudy Wind E 26 km/h" Output:

JSON
{
  "temperature": "29°C",
  "feels_like": "31°C", 
  "conditions": "Cloudy",
  "wind": "E 26 km/h"
}

Example 2: Daily Forecast Input: "Today 2/7 30° 23° Considerable cloudiness Night: Cloudy most of the time 25%" Output:

JSON
{
  "date": "2/7",
  "high": "30°",
  "low": "23°",
  "day_conditions": "Considerable cloudiness",
  "night_conditions": "Cloudy most of the time",
  "precipitation_chance": "25%"
}

Example 3: Hourly Data Input: "7 PM 26° 7% 8 PM 25° 7% 9 PM 25° 7%" Output:

JSON
[
  {"time": "7 PM", "temp": "26°", "precip_chance": "7%"},
  {"time": "8 PM", "temp": "25°", "precip_chance": "7%"},
  {"time": "9 PM", "temp": "25°", "precip_chance": "7%"}
]

• Look for temperature patterns (numbers followed by ° symbol)
• Identify time markers (AM/PM, dates, "Today", "Tonight")
• Extract percentage values for precipitation probability
• Preserve original units (°C/°F, km/h/mph)
• Capture both day and night conditions when available
• Include air quality and UV index if present
• Note sunrise/sunset times for context

• Don't confuse "feels like" temperature with actual temperature
• Don't mix current conditions with forecast data
• Don't assume temperature units - extract as provided
• Don't ignore night conditions in daily forecasts
• Don't parse navigation elements or ads as weather data
• Don't convert units unless specifically requested