気象庁の過去気象データをPandasでスクレイピング

Pandasでスクレイピング

その他内容は上記記事を参考

日ごと

import time
import pandas as pd

# 札幌
prec_no = 14
block_no = 47412

# 期間　※年と月だけ利用
start = "2022/01/01"
end = "2022/04/30"

# 期間中の月初を作成
dt_range = pd.date_range(start, end, freq="MS")

dfs = []

for dt in dt_range:

    url = f"https://www.data.jma.go.jp/obd/stats/etrn/view/daily_a1.php?prec_no={prec_no}&block_no={block_no}&year={year}&month={month}&day=&view="

    tmp = pd.read_html(url)[0]

    # 列名の重複除去して結合
    tmp.columns = [
        "_".join(sorted(set(col), key=col.index)) for col in tmp.columns.values
    ]

    tmp["年"] = dt.year
    tmp["月"] = dt.month

    dfs.append(tmp)

    # 1秒待機
    time.sleep(1)

# 結合
df = pd.concat(dfs).reset_index(drop=True)

# CSVに変換
df.to_csv("result.csv", encoding="utf_8_sig")

10分ごと

import time
import pandas as pd

# 横浜
prec_no = 46
block_no = 47670

# 期間
start = "2021/06/22"
end = "2021/06/30"

# 期間作成
dt_range = pd.date_range(start, end, freq="D")

dfs = []

for dt in dt_range:

    url = f"https://www.data.jma.go.jp/stats/etrn/view/10min_a1.php?prec_no={prec_no}&block_no={block_no}&year={dt.year}&month={dt.month}&day={dt.day}&view="

    tmp = pd.read_html(url)[0]

    # 列名の重複除去して結合
    tmp.columns = [
        "_".join(sorted(set(col), key=col.index)) for col in tmp.columns.values
    ]

    tmp["year"] = dt.year
    tmp["month"] = dt.month
    tmp["day"] = dt.day

    dfs.append(tmp)

    # 1秒待機
    time.sleep(1)

# 結合
df = pd.concat(dfs).reset_index(drop=True)

df[["hour", "minute"]] = df["時分"].str.split(":", expand=True).astype(int)

df["date"] = pd.to_datetime(df[["year", "month", "day"]])
df["time"] = df[["hour", "minute"]].apply(
    lambda x: pd.Timedelta(hours=x.hour, minutes=x.minute), axis=1
)

df["時分"] = df["date"] + df["time"]

df.drop(["date", "time", "year", "month", "day", "hour", "minute"], axis=1, inplace=True)

df.rename(columns={"時分": "日時"}, inplace=True)

# CSVに変換
df.to_csv("result.csv", encoding="utf_8_sig")