愛媛県の県立高等学校学科別入学志願者数（全日制）のPDFをCSVに変換

ehime-c.esnet.ed.jp

令和３年度県立高等学校学科別入学志願者数（志願変更後） のPDFをCSVに変換する

!wget https://ehime-c.esnet.ed.jp/koukou/nyuusi/r02nyuusi/r02isi_ato/atozenniti.pdf -O data.pdf

import pdfplumber
import pandas as pd

with pdfplumber.open("data.pdf") as pdf:

    dfs = []

    for page in pdf.pages:

        table = page.extract_table()

        df_tmp = pd.DataFrame(table[1:], columns=table[0])

        dfs.append(df_tmp)

df = pd.concat(dfs)

df1 = df.iloc[:, :7]
df2 = df.iloc[:, 7:]

df = pd.concat([df1, df2]).dropna(thresh=5)

df.set_axis(
    [
        "学校名",
        "本校・分校",
        "学科名",
        "定員",
        "入学志願者数",
        "推薦",
        "倍率",
    ],
    axis="columns",
    inplace=True,
)

df = df.mask(df.isna())

df["学校名"] = df["学校名"].fillna(method="ffill")

df["定員"] = df["定員"].str.replace(",", "").astype(int)
df["入学志願者数"] = df["入学志願者数"].str.replace(",", "").astype(int)
df["推薦"] = df["推薦"].str.replace(",", "").astype(int)

df["倍率"] = df["倍率"].astype(float)

# グループ内で補完
df["本校・分校"] = df.groupby("学校名")["本校・分校"].ffill()

df.reset_index(drop=True, inplace=True)
df.index += 1

df.to_csv("data.csv")

docs.google.com