サンプルCSVデータ

投稿者: | 2022-06-12
0件のコメント

目次

『データサイエンス100本ノック構造化データ加工編ガイドブック』より

1.ローデータ

  • 急ぎ使いたい場合は、１と３を使えばよい
import numpy as np
import pandas as pd
import pathlib
import re

url_customer = "https://miwadaice.info/sample_data/100knocks_csv/customer.csv"
url_category = "https://miwadaice.info/sample_data/100knocks_csv/category.csv"
url_product  = "https://miwadaice.info/sample_data/100knocks_csv/product.csv"
url_receipt  = "https://miwadaice.info/sample_data/100knocks_csv/receipt.csv"
url_store    = "https://miwadaice.info/sample_data/100knocks_csv/store.csv"
url_geocode  = "https://miwadaice.info/sample_data/100knocks_csv/geocode.csv"

2.型変換なしver

# 型変換なしバージョン
df_customer = pd.read_csv(url_customer)
df_category = pd.read_csv(url_category)
df_product = pd.read_csv(url_product)
df_receipt = pd.read_csv(url_receipt)
df_store = pd.read_csv(url_store)
df_geocode = pd.read_csv(url_geocode)

3.型変換ありver

  • dtype の辞書をつくりたいときは
  • {k: “” for k in df.columns.tolist()} で箱をつくっておくと楽
# 型指定ありバージョン
df_customer = pd.read_csv(url_customer,
                          dtype = {'customer_id': 'str',
                                 'customer_name': 'str',
                                 'gender_cd': 'str',
                                 'gender': 'str',
                                 # 'birth_day': '',
                                 'age': np.int64,
                                 'postal_cd': 'str',
                                 'address': 'str',
                                 'application_store_cd': 'str',
                                 'application_date': 'str',
                                 'status_cd': 'str'},
                            parse_dates=['birth_day']
                         )

df_customer['birth_day'] = df_customer['birth_day'].dt.date

df_category = pd.read_csv(url_category, dtype=str)

df_product = pd.read_csv(url_product,
                        dtype = {'product_cd': 'str',
                                 'category_major_cd': 'str',
                                 'category_medium_cd': 'str',
                                 'category_small_cd': 'str',
                                 'unit_price': np.float64,
                                 'unit_cost': np.float64}
                        )

df_receipt = pd.read_csv(url_receipt,
                        dtype = {'sales_ymd': np.int64,
                                 'sales_epoch': np.int64,
                                 'store_cd': 'str',
                                 'receipt_no': np.int64,
                                 'receipt_sub_no': np.int64,
                                 'customer_id': 'str',
                                 'product_cd': 'str',
                                 'quantity': np.int64,
                                 'amount': np.int64}
                        )

df_store = pd.read_csv(url_store,
                      dtype = {'store_cd': 'str',
                             'store_name': 'str',
                             'prefecture_cd': 'str',
                             'prefecture': 'str',
                             'address': 'str',
                             'address_kana': 'str',
                             'tel_no': 'str',
                             'longitude': np.float64,
                             'latitude': np.float64,
                             'floor_area': np.float64,
                          }
                      )

df_geocode = pd.read_csv(url_geocode,
                        dtype = {'postal_cd': 'str',
                                 'prefecture': 'str',
                                 'city': 'str',
                                 'town': 'str',
                                 'street': 'str',
                                 'address': 'str',
                                 'full_address': 'str',
                                 'longitude': np.float64,
                                 'latitude': np.float64}
                        )
import pandas as pd
import pathlib
import re

url_list = [url_customer, url_category, url_product, url_receipt, url_store, url_geocode]

# urlの末尾CSV名（拡張子除く）を取り出すregexパターン
p = re.compile(r"(\w*)(?:\.csv)$")

# サンプルデータを表示する
for url in url_list:
    df = pd.read_csv(url)
    csv_name = p.search(url).group(1)
    print(f'\n■ df_{csv_name}')
    display(df.head())