Extracting Flight Data from Email

Guide for identifying actual flight bookings and extracting structured information from Gmail.

For Gmail search query patterns, see Gmail Search Patterns (Travel > Flights section).

Identifying Flight Confirmations

Not every email with "flight" in the subject is a flight confirmation. Use inference judgment to filter.

Actual flight confirmations contain:

• Flight numbers (e.g., "UA 123", "BA456")
• PNR/confirmation codes (6 alphanumeric characters)
• Specific departure/arrival times and airports
• Airline name or logo
• Passenger name matching the user

Exclude:

• Hotel-only bookings (no flight info)
• Car rental confirmations
• Flight deal newsletters and promotions
• Price alerts and fare tracking emails
• Generic "thank you for booking" without flight details

Extracting Flight Details

For each confirmed flight email, extract:

Field	What to Look For	Example
Departure City	Airport code (3 letters) or city name near "from", "depart"	SFO, San Francisco
Arrival City	Airport code or city name near "to", "arrive"	CDG, Paris
Departure Date	Date near departure info, format varies	Jan 15, 2024, 2024-01-15
Departure Time	Time in 12h or 24h format	10:30 AM, 14:45
Arrival Date	Often same as departure, check for overnight	Jan 16, 2024
Arrival Time	Time at destination (local time)	07:00
Airline	Carrier name or 2-letter code	United, UA
Flight Number	Alphanumeric, often after airline	UA 837
Confirmation Code	6-character PNR	ABC123

Airport to Country Mapping

Common airport codes and their countries:

North America: JFK/EWR/LGA (USA-NY), LAX/SFO (USA-CA), ORD (USA-IL), YYZ (Canada), MEX (Mexico)

Europe: LHR/LGW (UK), CDG/ORY (France), FRA/MUC (Germany), AMS (Netherlands), FCO (Italy), MAD/BCN (Spain), ZRH (Switzerland)

Asia: NRT/HND (Japan), ICN (South Korea), PEK/PVG (China), SIN (Singapore), BKK (Thailand), HKG (Hong Kong)

Middle East: DXB (UAE), DOH (Qatar), TLV (Israel)

Oceania: SYD/MEL (Australia), AKL (New Zealand)

For unlisted airports, use the city name to determine country.

Handling Ambiguity

Multiple segments: A single confirmation may have multiple flights (outbound + return, or connections). Extract each segment separately.

Missing data: If departure time is missing but date is known, record the date. If airport code is ambiguous, prefer the full city name.

Duplicates: The same flight may appear multiple times (confirmation + reminder + check-in). Deduplicate by: same date + same route + same flight number.

Output Format

Write extracted flights as JSON array:

[
  {
    "source": "email",
    "confirmationCode": "ABC123",
    "airline": "United",
    "flightNumber": "UA 837",
    "departureCity": "San Francisco",
    "departureAirport": "SFO",
    "departureCountry": "USA",
    "departureDate": "2024-01-15",
    "departureTime": "10:30",
    "arrivalCity": "Paris",
    "arrivalAirport": "CDG",
    "arrivalCountry": "France",
    "arrivalDate": "2024-01-16",
    "arrivalTime": "07:00"
  }
]

Sort by departure date ascending. Include all fields even if some are null.