如何使用 TalorData 抓取 Google Flights 搜尋數據

Google Flights search data 適合用來取得即時航班資訊，例如路線、航空公司、票價、中轉、飛行時間、出發時間、抵達時間、booking options 和 fare changes。

對旅遊平台、OTA、價格監控團隊、市場研究團隊和 AI travel agents 來說，這些資料可以支撐：

重點是結構化。瀏覽器頁面不容易保存、比較或餵給系統。JSON data 更適合用於 dashboards、alerts、reports 和 AI workflows。

這篇指南會建立一個 Python workflow：透過 TalorData 發送 Google Flights request，解析 response，提取航班資料，最後匯出乾淨表格。

基本流程：

Route + date + passenger settings
   ↓
TalorData SERP API
   ↓
Google Flights data in JSON
   ↓
Parse flights, prices, airlines, stops, duration
   ↓
Save to CSV or database
   ↓
Use for monitoring, dashboards, or AI agents

可以收集哪些航班資料？

Google Flights workflow 最常用的欄位通常是 route、date、schedule、airline、price 和 booking context。結構化航班 response 通常可包含 departure and arrival airports、flight duration、layovers、carbon emissions、prices 和 price insights。

做 price monitoring 時，timestamp 不是可有可無。今天收集的 fare，明天可能就不一樣。

核心 Google Flights 參數

Google Flights request 通常從 route、date、passenger、localization 和 output settings 開始。Airport ID 通常是三位 IATA airport code，例如 JFK、LAX、LHR、SIN，有些 API 也支援 location identifiers，用於更寬泛的城市或地區搜尋。

One-way search 的核心是 departure_id、arrival_id 和 outbound_date。Round trip 需要加入 return_date。Multi-city itinerary 則使用 multi-city payload。

基本請求結構

Endpoint 和 authentication format 請以你的 TalorData account 為準。下面是簡單 one-way flight search 的 request shape。

{
  "engine": "google_flights",
  "departure_id": "JFK",
  "arrival_id": "LAX",
  "outbound_date": "2026-08-15",
  "flight_type": "one_way",
  "adults": 1,
  "currency": "USD",
  "hl": "en",
  "gl": "us",
  "no_cache": true
}

Round trip：

{
  "engine": "google_flights",
  "departure_id": "JFK",
  "arrival_id": "LAX",
  "outbound_date": "2026-08-15",
  "return_date": "2026-08-22",
  "flight_type": "round_trip",
  "adults": 1,
  "currency": "USD",
  "hl": "en",
  "gl": "us",
  "no_cache": true
}

Step 1：設定 API key

將 TalorData API key 和 endpoint 存成環境變數。

export TALORDATA_API_KEY="your_api_key_here"
export TALORDATA_SERP_ENDPOINT="your_talordata_serp_endpoint_here"

Windows PowerShell：

setx TALORDATA_API_KEY "your_api_key_here"
setx TALORDATA_SERP_ENDPOINT "your_talordata_serp_endpoint_here"

安裝 dependencies：

pip install requests pandas

Step 2：建立可重用 Python client

這個 function 會發送 Google Flights request 並返回 JSON。

import os
import requests
from typing import Any, Dict, Optional


TALORDATA_API_KEY = os.getenv("TALORDATA_API_KEY")
TALORDATA_SERP_ENDPOINT = os.getenv("TALORDATA_SERP_ENDPOINT")


def fetch_google_flights(
    departure_id: str,
    arrival_id: str,
    outbound_date: str,
    return_date: Optional[str] = None,
    flight_type: str = "one_way",
    adults: int = 1,
    children: int = 0,
    currency: str = "USD",
    hl: str = "en",
    gl: str = "us",
    travel_class: str = "economy",
    no_cache: bool = True,
    extra_params: Optional[Dict[str, Any]] = None
) -> Dict[str, Any]:
    """
    Fetch Google Flights search data with TalorData SERP API.
    """
    if not TALORDATA_API_KEY:
        raise RuntimeError("Missing TALORDATA_API_KEY environment variable.")

    if not TALORDATA_SERP_ENDPOINT:
        raise RuntimeError("Missing TALORDATA_SERP_ENDPOINT environment variable.")

    payload: Dict[str, Any] = {
        "engine": "google_flights",
        "departure_id": departure_id,
        "arrival_id": arrival_id,
        "outbound_date": outbound_date,
        "flight_type": flight_type,
        "adults": adults,
        "children": children,
        "currency": currency,
        "hl": hl,
        "gl": gl,
        "travel_class": travel_class,
        "no_cache": no_cache
    }

    if return_date:
        payload["return_date"] = return_date

    if extra_params:
        payload.update(extra_params)

    response = requests.post(
        TALORDATA_SERP_ENDPOINT,
        json=payload,
        headers={
            "Authorization": f"Bearer {TALORDATA_API_KEY}",
            "Content-Type": "application/json"
        },
        timeout=45
    )

    response.raise_for_status()
    return response.json()

基本執行：

if __name__ == "__main__":
    data = fetch_google_flights(
        departure_id="JFK",
        arrival_id="LAX",
        outbound_date="2026-08-15",
        flight_type="one_way",
        adults=1,
        currency="USD"
    )

    print(data)

Step 3：檢查 raw response

建立 parser 前，先檢查第一份 response。

import json

print(json.dumps(data, indent=2, ensure_ascii=False)[:6000])

Google Flights-style response 常見容器包括：

best_flights
other_flights
flights
price_insights
airports
search_parameters

實際結構可能因 request type、route 和 result mode 而不同。不要太早把 parser 刻成石碑。先用手電筒看 JSON 洞穴牆。

Step 4：解析 flight results

Google Flights response 常會把結果分成 best_flights 和 other_flights。每個 itinerary 可能包含一個或多個 flight segments。

from typing import Any, Dict, List


def get_flight_groups(data: Dict[str, Any]) -> List[Dict[str, Any]]:
    groups = []

    for group_name in ["best_flights", "other_flights", "flights"]:
        value = data.get(group_name)

        if isinstance(value, list):
            for item in value:
                item["_group"] = group_name
                groups.append(item)

    return groups


def parse_google_flights(data: Dict[str, Any]) -> List[Dict[str, Any]]:
    itineraries = get_flight_groups(data)
    rows = []

    for position, itinerary in enumerate(itineraries, start=1):
        segments = itinerary.get("flights", [])

        first_segment = segments[0] if segments else {}
        last_segment = segments[-1] if segments else {}

        departure_airport = first_segment.get("departure_airport", {})
        arrival_airport = last_segment.get("arrival_airport", {})

        airlines = sorted({
            segment.get("airline")
            for segment in segments
            if segment.get("airline")
        })

        flight_numbers = [
            segment.get("flight_number")
            for segment in segments
            if segment.get("flight_number")
        ]

        rows.append({
            "position": position,
            "group": itinerary.get("_group"),
            "airlines": ", ".join(airlines),
            "flight_numbers": ", ".join(flight_numbers),
            "departure_airport_name": departure_airport.get("name"),
            "departure_airport_id": departure_airport.get("id"),
            "departure_time": departure_airport.get("time"),
            "arrival_airport_name": arrival_airport.get("name"),
            "arrival_airport_id": arrival_airport.get("id"),
            "arrival_time": arrival_airport.get("time"),
            "duration": itinerary.get("total_duration") or itinerary.get("duration"),
            "stops": max(len(segments) - 1, 0),
            "price": itinerary.get("price"),
            "extracted_price": itinerary.get("extracted_price"),
            "currency": itinerary.get("currency"),
            "booking_token": itinerary.get("booking_token"),
            "departure_token": itinerary.get("departure_token"),
            "carbon_emissions": itinerary.get("carbon_emissions"),
            "layovers": itinerary.get("layovers"),
        })

    return rows

使用方式：

data = fetch_google_flights(
    departure_id="JFK",
    arrival_id="LAX",
    outbound_date="2026-08-15"
)

rows = parse_google_flights(data)

for row in rows[:5]:
    print(row["position"], row["airlines"], row["price"], row["stops"])

Step 5：保存成 CSV

import pandas as pd


data = fetch_google_flights(
    departure_id="JFK",
    arrival_id="LAX",
    outbound_date="2026-08-15",
    flight_type="one_way",
    adults=1,
    currency="USD"
)

rows = parse_google_flights(data)

df = pd.DataFrame(rows)
df.to_csv("google_flights_results.csv", index=False)

print(df.head())

CSV 可能長這樣：

Step 6：加入 search context

沒有 search context 的 flight result 很難比較。每次都要保存 route、date、passenger、currency、language、country 和 timestamp。

from datetime import datetime, timezone


def add_search_context(
    rows: List[Dict[str, Any]],
    departure_id: str,
    arrival_id: str,
    outbound_date: str,
    return_date: Optional[str],
    flight_type: str,
    adults: int,
    currency: str,
    hl: str,
    gl: str
) -> List[Dict[str, Any]]:
    collected_at = datetime.now(timezone.utc).isoformat()

    for row in rows:
        row["departure_id"] = departure_id
        row["arrival_id"] = arrival_id
        row["outbound_date"] = outbound_date
        row["return_date"] = return_date
        row["flight_type"] = flight_type
        row["adults"] = adults
        row["currency"] = currency
        row["hl"] = hl
        row["gl"] = gl
        row["collected_at"] = collected_at

    return rows

Step 7：追蹤價格變化

Fare monitoring 通常關心每條 route 和 date 的最低價格。

df["price_num"] = pd.to_numeric(df["extracted_price"], errors="coerce")

cheapest = df.sort_values(
    by=["price_num", "stops"],
    ascending=[True, True]
).head(10)

cheapest.to_csv("cheapest_google_flights.csv", index=False)

print(cheapest[[
    "position",
    "airlines",
    "departure_airport_id",
    "arrival_airport_id",
    "stops",
    "duration",
    "price",
    "price_num"
]])

可追蹤的 signals：

做 recurring jobs 時，不要覆蓋舊資料。應該 append 到 database 或 daily CSV。

Step 8：監控多條路線

routes = [
    {"departure_id": "JFK", "arrival_id": "LAX"},
    {"departure_id": "SFO", "arrival_id": "SEA"},
    {"departure_id": "LHR", "arrival_id": "CDG"},
]

all_rows = []

for route in routes:
    data = fetch_google_flights(
        departure_id=route["departure_id"],
        arrival_id=route["arrival_id"],
        outbound_date="2026-08-15",
        flight_type="one_way",
        adults=1,
        currency="USD"
    )

    rows = parse_google_flights(data)

    for row in rows:
        row["departure_id"] = route["departure_id"]
        row["arrival_id"] = route["arrival_id"]
        row["outbound_date"] = "2026-08-15"

    all_rows.extend(rows)

pd.DataFrame(all_rows).to_csv("multi_route_google_flights.csv", index=False)

適合：

Step 9：處理 round trips

Round-trip data 可能更複雜，因為 outbound selection 可能決定 return options。

簡單 round-trip request：

data = fetch_google_flights(
    departure_id="JFK",
    arrival_id="LAX",
    outbound_date="2026-08-15",
    return_date="2026-08-22",
    flight_type="round_trip",
    adults=1,
    currency="USD"
)

如果 response 返回 departure_token，你可以用第二次 request 取得 return-flight options。Google Flights-style API 常用 departure token 或 booking token 從初始結果繼續查下一段或 booking options。

Step 10：Multi-city flight search

Multi-city itinerary 使用多段 flight segments。

multi_city_json = [
    {
        "departure_id": "JFK",
        "arrival_id": "LHR",
        "date": "2026-08-15"
    },
    {
        "departure_id": "LHR",
        "arrival_id": "CDG",
        "date": "2026-08-20"
    },
    {
        "departure_id": "CDG",
        "arrival_id": "JFK",
        "date": "2026-08-25"
    }
]

data = fetch_google_flights(
    departure_id="",
    arrival_id="",
    outbound_date="",
    flight_type="multi_city",
    adults=1,
    currency="USD",
    extra_params={
        "multi_city_json": multi_city_json
    }
)

Multi-city 適合：

常見錯誤

不保存 timestamp

航班價格會變。每次 snapshot 都要有 collection time。

比較不同乘客設定

一位成人和兩位成人可能返回不同總價。要保存 passenger settings。

混用 currencies

比較前要保存 currency，並做價格標準化。

只看 price

低價但中轉兩次，未必比稍貴的直飛更好。

忽略 route context

JFK → LAX 和 NYC → Los Angeles area 不是同一個搜尋。

覆蓋舊資料

做 price monitoring 時，應該 append snapshots，而不是替換舊資料。

結語

使用 TalorData 抓取 Google Flights search data，本質上是把航班搜尋轉成結構化、可重複的資料流程。開始免費試用Google Flights API>>

先從 route 和 date parameters 開始：departure_id、arrival_id、outbound_date，需要 round trip 時加 return_date。再加入 passenger count、currency、language、country、travel class 和 filters。接著解析 airline、flight number、price、duration、stops、airports、times 和 booking tokens。

資料一旦結構化，就可以建立 price alerts、fare dashboards、airline visibility reports、travel market analysis 和 AI travel agents。

Google Flights data 像一塊一直跳動的出發看板。好的 API workflow，會把它變成系統能準時登機的資料集。