Posted 2022-03-27Updated 2022-03-26 HS Python27 minutes read (About 4116 words)

Pandas 입문 2

groupby 집중 연습

라이브러리 불러오기

pandas 라이브러리 불러오고, supermarket_sales.csv 파일 불러오기
참고로 미얀마 자료. 얀곤이 옛 수도, 네피도는 현 수도

1 2	import pandas as pd print(pd.__version__)

1.3.5

1 2	from google.colab import drive drive.mount('/content/drive') # "/content/drive" 도 가능

Mounted at /content/drive

1
2
3

DATA_PATH = '/content/drive/MyDrive/Colab Notebooks/data/supermarket_sales.csv'
sales = pd.read_csv(DATA_PATH) # 객체를 대문자로 썼는데, 개발자들이 좋아하는 방법이다. 눈에 띄는 걸 좋아한다. 
sales

	Invoice ID	Branch	City	Customer type	Gender	Product line	Unit price	Quantity	Date	Time	Payment
0	750-67-8428	A	Yangon	Member	Female	Health and beauty	74.69	7	1/5/2019	13:08	Ewallet
1	226-31-3081	C	Naypyitaw	Normal	Female	Electronic accessories	15.28	5	3/8/2019	10:29	Cash
2	631-41-3108	A	Yangon	Normal	Male	Home and lifestyle	46.33	7	3/3/2019	13:23	Credit card
3	123-19-1176	A	Yangon	Member	Male	Health and beauty	58.22	8	1/27/2019	20:33	Ewallet
4	373-73-7910	A	Yangon	Normal	Male	Sports and travel	86.31	7	2/8/2019	10:37	Ewallet
...	...	...	...	...	...	...	...	...	...	...	...
995	233-67-5758	C	Naypyitaw	Normal	Male	Health and beauty	40.35	1	1/29/2019	13:46	Ewallet
996	303-96-2227	B	Mandalay	Normal	Female	Home and lifestyle	97.38	10	3/2/2019	17:16	Ewallet
997	727-02-1313	A	Yangon	Member	Male	Food and beverages	31.84	1	2/9/2019	13:22	Cash
998	347-56-2442	A	Yangon	Normal	Male	Home and lifestyle	65.82	1	2/22/2019	15:33	Cash
999	849-09-3807	A	Yangon	Member	Female	Fashion accessories	88.34	7	2/18/2019	13:28	Cash

1000 rows × 11 columns

  <script>
    const buttonEl =
      document.querySelector('#df-2c107087-f544-4f62-bedd-767a947c87b5 button.colab-df-convert');
    buttonEl.style.display =
      google.colab.kernel.accessAllowed ? 'block' : 'none';

    async function convertToInteractive(key) {
      const element = document.querySelector('#df-2c107087-f544-4f62-bedd-767a947c87b5');
      const dataTable =
        await google.colab.kernel.invokeFunction('convertToInteractive',
                                                 [key], {});
      if (!dataTable) return;

      const docLinkHtml = 'Like what you see? Visit the ' +
        '<a target="_blank" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'
        + ' to learn more about interactive tables.';
      element.innerHTML = '';
      dataTable['output_type'] = 'display_data';
      await google.colab.output.renderOutput(dataTable, element);
      const docLink = document.createElement('div');
      docLink.innerHTML = docLinkHtml;
      element.appendChild(docLink);
    }
  </script>
</div>

역시 인포부터 알아본다.

1	sales.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1000 entries, 0 to 999
Data columns (total 11 columns):
 #   Column         Non-Null Count  Dtype  
---  ------         --------------  -----  
 0   Invoice ID     1000 non-null   object 
 1   Branch         1000 non-null   object 
 2   City           1000 non-null   object 
 3   Customer type  1000 non-null   object 
 4   Gender         1000 non-null   object 
 5   Product line   1000 non-null   object 
 6   Unit price     1000 non-null   float64
 7   Quantity       1000 non-null   int64  
 8   Date           1000 non-null   object 
 9   Time           1000 non-null   object 
 10  Payment        1000 non-null   object 
dtypes: float64(1), int64(1), object(9)
memory usage: 86.1+ KB

Group by

(동의어) 집계함수를 배운다.

1
2
3

sales['Invoice ID'].value_counts()
# 송장번호로 그룹화 하기는 어렵겠다...
# 적절한 기준을 찾을 때까지 노가다 해야한다.

750-67-8428    1
642-61-4706    1
816-72-8853    1
491-38-3499    1
322-02-2271    1
              ..
633-09-3463    1
374-17-3652    1
378-07-7001    1
433-75-6987    1
849-09-3807    1
Name: Invoice ID, Length: 1000, dtype: int64

1 2	sales.groupby('Product line')['Quantity'].sum() # 어떤 칼럼으로 그룹을 지을지 고민해야한다.

Product line
Electronic accessories    971
Fashion accessories       902
Food and beverages        952
Health and beauty         854
Home and lifestyle        911
Sports and travel         920
Name: Quantity, dtype: int64

1	sales.groupby(['Product line', 'Branch', 'Payment'])['Quantity'].sum()

Product line            Branch  Payment    
Electronic accessories  A       Cash            85
                                Credit card    121
                                Ewallet        116
                        B       Cash           134
                                Credit card     83
                                Ewallet         99
                        C       Cash           179
                                Credit card     58
                                Ewallet         96
Fashion accessories     A       Cash            60
                                Credit card     85
                                Ewallet        118
                        B       Cash           110
                                Credit card    100
                                Ewallet         87
                        C       Cash           110
                                Credit card    108
                                Ewallet        124
Food and beverages      A       Cash            78
                                Credit card    114
                                Ewallet        121
                        B       Cash            41
                                Credit card    138
                                Ewallet         91
                        C       Cash           176
                                Credit card     83
                                Ewallet        110
Health and beauty       A       Cash            98
                                Credit card     73
                                Ewallet         86
                        B       Cash           113
                                Credit card     97
                                Ewallet        110
                        C       Cash            82
                                Credit card    104
                                Ewallet         91
Home and lifestyle      A       Cash           139
                                Credit card     85
                                Ewallet        147
                        B       Cash            94
                                Credit card     93
                                Ewallet        108
                        C       Cash            73
                                Credit card     81
                                Ewallet         91
Sports and travel       A       Cash           112
                                Credit card    102
                                Ewallet        119
                        B       Cash           136
                                Credit card     88
                                Ewallet         98
                        C       Cash            76
                                Credit card    109
                                Ewallet         80
Name: Quantity, dtype: int64

여기서 시리즈타입과 데이터프레임타입 구분이 중요하다.
위의 표는 무슨 타입일까?

1	print(type(sales.groupby(['Product line', 'Branch', 'Payment'])['Quantity'].sum()))

<class 'pandas.core.series.Series'>

이게 시리즈 타입이라고?? 시리즈는 인덱스 하나 + 칼럼 하나로 구성된 것이 아니었나…
Product line, Branch, Payment 세 개가 함께 인덱스란 소리다.
“Multi index”
가장 오른쪽에 나오는 숫자열이 한 개의 칼럼에 해당

시리즈 타입을 데이터프레임 타입으로 바꾸고 싶어

구글링 필요.
as_index = False

1	sales.groupby(['Product line', 'Branch', 'Payment'], as_index=False)['Quantity'].sum()

	Product line	Branch	Payment	Quantity
0	Electronic accessories	A	Cash	85
1	Electronic accessories	A	Credit card	121
2	Electronic accessories	A	Ewallet	116
3	Electronic accessories	B	Cash	134
4	Electronic accessories	B	Credit card	83
5	Electronic accessories	B	Ewallet	99
6	Electronic accessories	C	Cash	179
7	Electronic accessories	C	Credit card	58
8	Electronic accessories	C	Ewallet	96
9	Fashion accessories	A	Cash	60
10	Fashion accessories	A	Credit card	85
11	Fashion accessories	A	Ewallet	118
12	Fashion accessories	B	Cash	110
13	Fashion accessories	B	Credit card	100
14	Fashion accessories	B	Ewallet	87
15	Fashion accessories	C	Cash	110
16	Fashion accessories	C	Credit card	108
17	Fashion accessories	C	Ewallet	124
18	Food and beverages	A	Cash	78
19	Food and beverages	A	Credit card	114
20	Food and beverages	A	Ewallet	121
21	Food and beverages	B	Cash	41
22	Food and beverages	B	Credit card	138
23	Food and beverages	B	Ewallet	91
24	Food and beverages	C	Cash	176
25	Food and beverages	C	Credit card	83
26	Food and beverages	C	Ewallet	110
27	Health and beauty	A	Cash	98
28	Health and beauty	A	Credit card	73
29	Health and beauty	A	Ewallet	86
30	Health and beauty	B	Cash	113
31	Health and beauty	B	Credit card	97
32	Health and beauty	B	Ewallet	110
33	Health and beauty	C	Cash	82
34	Health and beauty	C	Credit card	104
35	Health and beauty	C	Ewallet	91
36	Home and lifestyle	A	Cash	139
37	Home and lifestyle	A	Credit card	85
38	Home and lifestyle	A	Ewallet	147
39	Home and lifestyle	B	Cash	94
40	Home and lifestyle	B	Credit card	93
41	Home and lifestyle	B	Ewallet	108
42	Home and lifestyle	C	Cash	73
43	Home and lifestyle	C	Credit card	81
44	Home and lifestyle	C	Ewallet	91
45	Sports and travel	A	Cash	112
46	Sports and travel	A	Credit card	102
47	Sports and travel	A	Ewallet	119
48	Sports and travel	B	Cash	136
49	Sports and travel	B	Credit card	88
50	Sports and travel	B	Ewallet	98
51	Sports and travel	C	Cash	76
52	Sports and travel	C	Credit card	109
53	Sports and travel	C	Ewallet	80

  <script>
    const buttonEl =
      document.querySelector('#df-8d1795bd-e055-460d-8f14-0c4e1c3f00d5 button.colab-df-convert');
    buttonEl.style.display =
      google.colab.kernel.accessAllowed ? 'block' : 'none';

    async function convertToInteractive(key) {
      const element = document.querySelector('#df-8d1795bd-e055-460d-8f14-0c4e1c3f00d5');
      const dataTable =
        await google.colab.kernel.invokeFunction('convertToInteractive',
                                                 [key], {});
      if (!dataTable) return;

      const docLinkHtml = 'Like what you see? Visit the ' +
        '<a target="_blank" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'
        + ' to learn more about interactive tables.';
      element.innerHTML = '';
      dataTable['output_type'] = 'display_data';
      await google.colab.output.renderOutput(dataTable, element);
      const docLink = document.createElement('div');
      docLink.innerHTML = docLinkHtml;
      element.appendChild(docLink);
    }
  </script>
</div>

1	print(type(sales.groupby(['Product line', 'Branch', 'Payment'], as_index=False)['Quantity'].sum()))

<class 'pandas.core.frame.DataFrame'>

결측치 다루기

결측치 데이터 생성

임의로 만들기
대부분 딕셔너리 형태

import pandas as pd
import numpy as np

dict_01 = {
    'Score_A' : [80,90,np.nan,80],
    'Score_B' : [30,45,np.nan, np.nan],
    'Score_C' : [np.nan, 50,80,90]    
}
df = pd.DataFrame(dict_01)
df

	Score_A	Score_B	Score_C
0	80.0	30.0	NaN
1	90.0	45.0	50.0
2	NaN	NaN	80.0
3	80.0	NaN	90.0

  <script>
    const buttonEl =
      document.querySelector('#df-cf88f083-c896-4734-b449-7286820b6a16 button.colab-df-convert');
    buttonEl.style.display =
      google.colab.kernel.accessAllowed ? 'block' : 'none';

    async function convertToInteractive(key) {
      const element = document.querySelector('#df-cf88f083-c896-4734-b449-7286820b6a16');
      const dataTable =
        await google.colab.kernel.invokeFunction('convertToInteractive',
                                                 [key], {});
      if (!dataTable) return;

      const docLinkHtml = 'Like what you see? Visit the ' +
        '<a target="_blank" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'
        + ' to learn more about interactive tables.';
      element.innerHTML = '';
      dataTable['output_type'] = 'display_data';
      await google.colab.output.renderOutput(dataTable, element);
      const docLink = document.createElement('div');
      docLink.innerHTML = docLinkHtml;
      element.appendChild(docLink);
    }
  </script>
</div>

1	df.isnull().sum() # 각 컬럼의 결측치 개수

Score_A    1
Score_B    2
Score_C    1
dtype: int64

True = 숫자 1로 인식
False = 숫자 0으로 인식

1
2

type(df.isnull().sum()) # 중간중간 타입을 확인하는 이유는 시리즈와 데이터프레임 메서드가 다르니까 그런것이다.
# 표 보다는 그래프로 보여주면 더 인식하기 쉽다.

pandas.core.series.Series

1 2	df.fillna("입력값") # "입력값"으로 채우기 # df.fillna("0") -> 0으로 채우기

	Score_A	Score_B	Score_C
0	80.0	30.0	입력값
1	90.0	45.0	50.0
2	입력값	입력값	80.0
3	80.0	입력값	90.0

  <script>
    const buttonEl =
      document.querySelector('#df-a445638e-d7a2-4313-8ad8-ec710da9a067 button.colab-df-convert');
    buttonEl.style.display =
      google.colab.kernel.accessAllowed ? 'block' : 'none';

    async function convertToInteractive(key) {
      const element = document.querySelector('#df-a445638e-d7a2-4313-8ad8-ec710da9a067');
      const dataTable =
        await google.colab.kernel.invokeFunction('convertToInteractive',
                                                 [key], {});
      if (!dataTable) return;

      const docLinkHtml = 'Like what you see? Visit the ' +
        '<a target="_blank" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'
        + ' to learn more about interactive tables.';
      element.innerHTML = '';
      dataTable['output_type'] = 'display_data';
      await google.colab.output.renderOutput(dataTable, element);
      const docLink = document.createElement('div');
      docLink.innerHTML = docLinkHtml;
      element.appendChild(docLink);
    }
  </script>
</div>

1	df.fillna(method="pad") # 결측치 바로 위에 있는 값으로 채우기

	Score_A	Score_B	Score_C
0	80.0	30.0	NaN
1	90.0	45.0	50.0
2	90.0	45.0	80.0
3	80.0	45.0	90.0

  <script>
    const buttonEl =
      document.querySelector('#df-8721f51c-2ef4-419b-8511-8e0a18fd252a button.colab-df-convert');
    buttonEl.style.display =
      google.colab.kernel.accessAllowed ? 'block' : 'none';

    async function convertToInteractive(key) {
      const element = document.querySelector('#df-8721f51c-2ef4-419b-8511-8e0a18fd252a');
      const dataTable =
        await google.colab.kernel.invokeFunction('convertToInteractive',
                                                 [key], {});
      if (!dataTable) return;

      const docLinkHtml = 'Like what you see? Visit the ' +
        '<a target="_blank" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'
        + ' to learn more about interactive tables.';
      element.innerHTML = '';
      dataTable['output_type'] = 'display_data';
      await google.colab.output.renderOutput(dataTable, element);
      const docLink = document.createElement('div');
      docLink.innerHTML = docLinkHtml;
      element.appendChild(docLink);
    }
  </script>
</div>

dict_01 = {
    "성별" : ["남자", "여자", np.nan, "남자"],
    "Salary" : [30, 45, 90, 70]
}
df = pd.DataFrame(dict_01)
df

	성별	Salary
0	남자	30
1	여자	45
2	NaN	90
3	남자	70

  <script>
    const buttonEl =
      document.querySelector('#df-c9256c94-14de-4167-a50f-4ab030334a2d button.colab-df-convert');
    buttonEl.style.display =
      google.colab.kernel.accessAllowed ? 'block' : 'none';

    async function convertToInteractive(key) {
      const element = document.querySelector('#df-c9256c94-14de-4167-a50f-4ab030334a2d');
      const dataTable =
        await google.colab.kernel.invokeFunction('convertToInteractive',
                                                 [key], {});
      if (!dataTable) return;

      const docLinkHtml = 'Like what you see? Visit the ' +
        '<a target="_blank" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'
        + ' to learn more about interactive tables.';
      element.innerHTML = '';
      dataTable['output_type'] = 'display_data';
      await google.colab.output.renderOutput(dataTable, element);
      const docLink = document.createElement('div');
      docLink.innerHTML = docLinkHtml;
      element.appendChild(docLink);
    }
  </script>
</div>

1	df['성별'].fillna("성별 없음") # 성별 칼럼에서만 결측치 원하는 값으로 채우기

0       남자
1       여자
2    성별 없음
3       남자
Name: 성별, dtype: object

결측치

–> 문자열 타입 / 숫자 타입 접근 방법이 다르다.
–> 문자열 (빈도 –> 가장 많이 나타나는 문자열 넣어주기!, 최빈값)
–> 숫자열 (평균, 최대, 최소, 중간, 이상치로 처리, 기타 등등..)

Score_D 결측치 없는 칼럼 추가

import pandas as pd
import numpy as np

dict_01 = {
    'Score_A' : [80,90,np.nan,80],
    'Score_B' : [30,45,np.nan, 60],
    'Score_C' : [np.nan, 50,80,90],
    'Score_D' : [50, 30, 80, 60]   
}
df = pd.DataFrame(dict_01)
df

	Score_A	Score_B	Score_C	Score_D
0	80.0	30.0	NaN	50
1	90.0	45.0	50.0	30
2	NaN	NaN	80.0	80
3	80.0	60.0	90.0	60

  <script>
    const buttonEl =
      document.querySelector('#df-e3909644-07fd-4398-9f80-787495b12866 button.colab-df-convert');
    buttonEl.style.display =
      google.colab.kernel.accessAllowed ? 'block' : 'none';

    async function convertToInteractive(key) {
      const element = document.querySelector('#df-e3909644-07fd-4398-9f80-787495b12866');
      const dataTable =
        await google.colab.kernel.invokeFunction('convertToInteractive',
                                                 [key], {});
      if (!dataTable) return;

      const docLinkHtml = 'Like what you see? Visit the ' +
        '<a target="_blank" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'
        + ' to learn more about interactive tables.';
      element.innerHTML = '';
      dataTable['output_type'] = 'display_data';
      await google.colab.output.renderOutput(dataTable, element);
      const docLink = document.createElement('div');
      docLink.innerHTML = docLinkHtml;
      element.appendChild(docLink);
    }
  </script>
</div>

인덱스1, 3(1행, 3행), D열은 결측치가 없다.

1 2	df.dropna(axis = 1) # axis = 1 은 컬럼기준 # axis = 0 은 인덱스 기준 # 결측치가 있는 컬럼은 없애주겠다.

	Score_D
0	50
1	30
2	80
3	60

  <script>
    const buttonEl =
      document.querySelector('#df-4e41ca0e-e4ca-40cc-a91c-14b180a756bf button.colab-df-convert');
    buttonEl.style.display =
      google.colab.kernel.accessAllowed ? 'block' : 'none';

    async function convertToInteractive(key) {
      const element = document.querySelector('#df-4e41ca0e-e4ca-40cc-a91c-14b180a756bf');
      const dataTable =
        await google.colab.kernel.invokeFunction('convertToInteractive',
                                                 [key], {});
      if (!dataTable) return;

      const docLinkHtml = 'Like what you see? Visit the ' +
        '<a target="_blank" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'
        + ' to learn more about interactive tables.';
      element.innerHTML = '';
      dataTable['output_type'] = 'display_data';
      await google.colab.output.renderOutput(dataTable, element);
      const docLink = document.createElement('div');
      docLink.innerHTML = docLinkHtml;
      element.appendChild(docLink);
    }
  </script>
</div>

1	df.dropna(axis = 0) # 결측치가 있는 인덱스(행)은 없애주겠다.

	Score_A	Score_B	Score_C	Score_D
1	90.0	45.0	50.0	30
3	80.0	60.0	90.0	60

  <script>
    const buttonEl =
      document.querySelector('#df-7d8b6c36-b63e-46c9-a499-6fd349891ef2 button.colab-df-convert');
    buttonEl.style.display =
      google.colab.kernel.accessAllowed ? 'block' : 'none';

    async function convertToInteractive(key) {
      const element = document.querySelector('#df-7d8b6c36-b63e-46c9-a499-6fd349891ef2');
      const dataTable =
        await google.colab.kernel.invokeFunction('convertToInteractive',
                                                 [key], {});
      if (!dataTable) return;

      const docLinkHtml = 'Like what you see? Visit the ' +
        '<a target="_blank" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'
        + ' to learn more about interactive tables.';
      element.innerHTML = '';
      dataTable['output_type'] = 'display_data';
      await google.colab.output.renderOutput(dataTable, element);
      const docLink = document.createElement('div');
      docLink.innerHTML = docLinkHtml;
      element.appendChild(docLink);
    }
  </script>
</div>

이상치

1
2

sales

	Invoice ID	Branch	City	Customer type	Gender	Product line	Unit price	Quantity	Date	Time	Payment
0	750-67-8428	A	Yangon	Member	Female	Health and beauty	74.69	7	1/5/2019	13:08	Ewallet
1	226-31-3081	C	Naypyitaw	Normal	Female	Electronic accessories	15.28	5	3/8/2019	10:29	Cash
2	631-41-3108	A	Yangon	Normal	Male	Home and lifestyle	46.33	7	3/3/2019	13:23	Credit card
3	123-19-1176	A	Yangon	Member	Male	Health and beauty	58.22	8	1/27/2019	20:33	Ewallet
4	373-73-7910	A	Yangon	Normal	Male	Sports and travel	86.31	7	2/8/2019	10:37	Ewallet
...	...	...	...	...	...	...	...	...	...	...	...
995	233-67-5758	C	Naypyitaw	Normal	Male	Health and beauty	40.35	1	1/29/2019	13:46	Ewallet
996	303-96-2227	B	Mandalay	Normal	Female	Home and lifestyle	97.38	10	3/2/2019	17:16	Ewallet
997	727-02-1313	A	Yangon	Member	Male	Food and beverages	31.84	1	2/9/2019	13:22	Cash
998	347-56-2442	A	Yangon	Normal	Male	Home and lifestyle	65.82	1	2/22/2019	15:33	Cash
999	849-09-3807	A	Yangon	Member	Female	Fashion accessories	88.34	7	2/18/2019	13:28	Cash

1000 rows × 11 columns

  <script>
    const buttonEl =
      document.querySelector('#df-eadd2087-3f9d-4047-a8f0-8d2fa7572222 button.colab-df-convert');
    buttonEl.style.display =
      google.colab.kernel.accessAllowed ? 'block' : 'none';

    async function convertToInteractive(key) {
      const element = document.querySelector('#df-eadd2087-3f9d-4047-a8f0-8d2fa7572222');
      const dataTable =
        await google.colab.kernel.invokeFunction('convertToInteractive',
                                                 [key], {});
      if (!dataTable) return;

      const docLinkHtml = 'Like what you see? Visit the ' +
        '<a target="_blank" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'
        + ' to learn more about interactive tables.';
      element.innerHTML = '';
      dataTable['output_type'] = 'display_data';
      await google.colab.output.renderOutput(dataTable, element);
      const docLink = document.createElement('div');
      docLink.innerHTML = docLinkHtml;
      element.appendChild(docLink);
    }
  </script>
</div>

일반적인 통계 공식은
IQR(Interquartile Range) - 박스플롯 - 사분위수
Q0(0), Q1(25%), Q2(50%), Q3(75%), Q4(100%)
IQR = Q3 - Q1
이상치의 하한 경계값 : Q1 - (1.5 * (Q3-Q1))
이상치의 상한 경계값 : Q3 + (1.5 * (Q3-Q1))
도메인(각 비즈니스 영역, 미래 일자리)에서 바라보는 이상치 기준(관습)
더 중요한 것은 관습이다.

1	sales[['Unit price']].describe()

	Unit price
count	1000.000000
mean	55.672130
std	26.494628
min	10.080000
25%	32.875000
50%	55.230000
75%	77.935000
max	99.960000

  <script>
    const buttonEl =
      document.querySelector('#df-c275e09e-666c-4e82-b6f2-72bbe50e96e4 button.colab-df-convert');
    buttonEl.style.display =
      google.colab.kernel.accessAllowed ? 'block' : 'none';

    async function convertToInteractive(key) {
      const element = document.querySelector('#df-c275e09e-666c-4e82-b6f2-72bbe50e96e4');
      const dataTable =
        await google.colab.kernel.invokeFunction('convertToInteractive',
                                                 [key], {});
      if (!dataTable) return;

      const docLinkHtml = 'Like what you see? Visit the ' +
        '<a target="_blank" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'
        + ' to learn more about interactive tables.';
      element.innerHTML = '';
      dataTable['output_type'] = 'display_data';
      await google.colab.output.renderOutput(dataTable, element);
      const docLink = document.createElement('div');
      docLink.innerHTML = docLinkHtml;
      element.appendChild(docLink);
    }
  </script>
</div>

Q1 = sales['Unit price'].quantile(0.25)
Q3 = sales['Unit price'].quantile(0.75)
outliers_Q1 = Q1 - 1.5 * (Q3-Q1)
outliers_Q1 # -34.~~ 값은 쓰기 어렵다. 이럴 경우에는 임의로 정해준다.
# 이상치의 하한값을 그냥 Q1으로 잡는다.
real_outliers_Q1 = (sales['Unit price'] < Q1)

# 이상치의 상한값을 마찬가지로 Q3로 잡는다.
real_outliers_Q3 = (sales['Unit price'] > Q3)

1	print(sales['Unit price'][~(real_outliers_Q1\|real_outliers_Q3)])

0      74.69
2      46.33
3      58.22
6      68.84
7      73.56
       ...  
991    76.60
992    58.03
994    60.95
995    40.35
998    65.82
Name: Unit price, Length: 500, dtype: float64

Pandas 입문 2

http://griotold.github.io/2022/03/27/pandas_tutorial_02/

Author

HS

Posted on

2022-03-27

Updated on

2022-03-26

Licensed under

You need to set install_url to use ShareThis. Please set it in _config.yml.

Afdian.net Alipay

Buy me a coffee Patreon

You forgot to set the business or currency_code for Paypal. Please set it in _config.yml.

Wechat

Comments

You forgot to set the shortname for Disqus. Please set it in _config.yml.

Pandas 입문 2

groupby 집중 연습

라이브러리 불러오기

Group by

시리즈 타입을 데이터프레임 타입으로 바꾸고 싶어

결측치 다루기

결측치 데이터 생성

Score_D 결측치 없는 칼럼 추가

이상치

Author

Posted on

Updated on

Licensed under

Like this article? Support the author with

Comments

Links

Categories

Recents

Archives

Subscribe for updates

follow.it