Pandas - DataFrame의 각 columns 분석을 위한 유틸리티 함수

2023-12-25 2023-12-25 917 words 5 minutes

Contents

DataFrame의 각 열이 어떤 데이터들을 가지고 있는지 간편하게 살펴보자.

pandas 로고

1. 핵심 요약

아래의 inspect_columns() 함수로 각 열의 데이터를 확인할 수 있음

def inspect_columns(df):
    df_describe = df.describe()
    df_result = pd.DataFrame(
        data={
            'dtype': df.dtypes,
            'unique': (df.nunique() == len(df)) + 0,  # 0: False, 1: True
            'nunique': df.nunique().map('{:,d}'.format),
            'null_count': df.isna().sum().map('{:,d}'.format),
            'null_pct': ((df.isnull().sum() / len(df)) * 100).map('{:,.2f}'.format),
            '|': '▮',
            'min': df_describe.loc['min'].map('{:,.2f}'.format),
            'mean': df_describe.loc['mean'].map('{:,.2f}'.format),
            'median': df_describe.loc['50%'].map('{:,.2f}'.format),
            'max': df_describe.loc['max'].map('{:,.2f}'.format),
            'std': df_describe.loc['std'].map('{:,.2f}'.format),
            '||': '▮',
            '1st_row': df.iloc[0],
            'random_row': df.iloc[np.random.randint(low=0, high=len(df))],
            'last_row': df.iloc[-1],
        },
        index=df.columns,
    )

    return df_result

2. 사용 예시

import numpy as np
import pandas as pd
import seaborn as sns

2.1. 유틸리티 함수

def inspect_columns(df):
    df_describe = df.describe()
    df_result = pd.DataFrame(
        data={
            'dtype': df.dtypes,
            'unique': (df.nunique() == len(df)) + 0,  # 0: False, 1: True
            'nunique': df.nunique().map('{:,d}'.format),
            'null_count': df.isna().sum().map('{:,d}'.format),
            'null_pct': ((df.isnull().sum() / len(df)) * 100).map('{:,.2f}'.format),
            '|': '▮',
            'min': df_describe.loc['min'].map('{:,.2f}'.format),
            'mean': df_describe.loc['mean'].map('{:,.2f}'.format),
            'median': df_describe.loc['50%'].map('{:,.2f}'.format),
            'max': df_describe.loc['max'].map('{:,.2f}'.format),
            'std': df_describe.loc['std'].map('{:,.2f}'.format),
            '||': '▮',
            '1st_row': df.iloc[0],
            'random_row': df.iloc[np.random.randint(low=0, high=len(df))],
            'last_row': df.iloc[-1],
        },
        index=df.columns,
    )

    return df_result

2.2. titanic

df_titanic = sns.load_dataset('titanic')
print(df_titanic.shape)
df_titanic.head()

(891, 15)

	survived	pclass	sex	age	sibsp	fare	embarked	class	who	adult_male	deck	embark_town	alive	alone
0	0	3	male	22.0	1	7.2500	S	Third	man	True	NaN	Southampton	no	False
1	1	1	female	38.0	1	71.2833	C	First	woman	False	C	Cherbourg	yes	False
2	1	3	female	26.0	0	7.9250	S	Third	woman	False	NaN	Southampton	yes	True
3	1	1	female	35.0	1	53.1000	S	First	woman	False	C	Southampton	yes	False
4	0	3	male	35.0	0	8.0500	S	Third	man	True	NaN	Southampton	no	True

inspect_columns(df_titanic)

	dtype	nunique	null_count	null_pct	\|	min	mean	median	max	std	\|\|	1st_row	random_row	last_row
survived	int64	2	0	0.00	▮	0.00	0.38	0.00	1.00	0.49	▮	0	0	0
pclass	int64	3	0	0.00	▮	1.00	2.31	3.00	3.00	0.84	▮	3	3	3
sex	object	2	0	0.00	▮	NaN	NaN	NaN	NaN	NaN	▮	male	male	male
age	float64	88	177	19.87	▮	0.42	29.70	28.00	80.00	14.53	▮	22.0	27.0	32.0
sibsp	int64	7	0	0.00	▮	0.00	0.52	0.00	8.00	1.10	▮	1	0	0
parch	int64	7	0	0.00	▮	0.00	0.38	0.00	6.00	0.81	▮	0	0	0
fare	float64	248	0	0.00	▮	0.00	32.20	14.45	512.33	49.69	▮	7.25	7.8958	7.75
embarked	object	3	2	0.22	▮	NaN	NaN	NaN	NaN	NaN	▮	S	S	Q
class	category	3	0	0.00	▮	NaN	NaN	NaN	NaN	NaN	▮	Third	Third	Third
who	object	3	0	0.00	▮	NaN	NaN	NaN	NaN	NaN	▮	man	man	man
adult_male	bool	2	0	0.00	▮	NaN	NaN	NaN	NaN	NaN	▮	True	True	True
deck	category	7	688	77.22	▮	NaN	NaN	NaN	NaN	NaN	▮	NaN	NaN	NaN
embark_town	object	3	2	0.22	▮	NaN	NaN	NaN	NaN	NaN	▮	Southampton	Southampton	Queenstown
alive	object	2	0	0.00	▮	NaN	NaN	NaN	NaN	NaN	▮	no	no	no
alone	bool	2	0	0.00	▮	NaN	NaN	NaN	NaN	NaN	▮	False	True	True

2.3. iris

df_iris = sns.load_dataset('iris')
print(df_iris.shape)
df_iris.head()

(150, 5)

	sepal_length	sepal_width	petal_length	petal_width	species
0	5.1	3.5	1.4	0.2	setosa
1	4.9	3.0	1.4	0.2	setosa
2	4.7	3.2	1.3	0.2	setosa
3	4.6	3.1	1.5	0.2	setosa
4	5.0	3.6	1.4	0.2	setosa

inspect_columns(df_iris)

	dtype	nunique	\|	min	mean	median	max	std	\|\|	1st_row	random_row	last_row
sepal_length	float64	35	▮	4.30	5.84	5.80	7.90	0.83	▮	5.1	6.8	5.9
sepal_width	float64	23	▮	2.00	3.06	3.00	4.40	0.44	▮	3.5	3.0	3.0
petal_length	float64	43	▮	1.00	3.76	4.35	6.90	1.77	▮	1.4	5.5	5.1
petal_width	float64	22	▮	0.10	1.20	1.30	2.50	0.76	▮	0.2	2.1	1.8
species	object	3	▮	NaN	NaN	NaN	NaN	NaN	▮	setosa	virginica	virginica

2.4. diamonds

df_diamonds = sns.load_dataset('diamonds')
print(df_diamonds.shape)
df_diamonds.head()

(53940, 10)

	carat	cut	color	clarity	depth	table	price	x	y	z
0	0.23	Ideal	E	SI2	61.5	55.0	326	3.95	3.98	2.43
1	0.21	Premium	E	SI1	59.8	61.0	326	3.89	3.84	2.31
2	0.23	Good	E	VS1	56.9	65.0	327	4.05	4.07	2.31
3	0.29	Premium	I	VS2	62.4	58.0	334	4.20	4.23	2.63
4	0.31	Good	J	SI2	63.3	58.0	335	4.34	4.35	2.75

inspect_columns(df_diamonds)

	dtype	nunique	\|	min	mean	median	max	std	\|\|	1st_row	random_row	last_row
carat	float64	273	▮	0.20	0.80	0.70	5.01	0.47	▮	0.23	0.71	0.75
cut	category	5	▮	NaN	NaN	NaN	NaN	NaN	▮	Ideal	Premium	Ideal
color	category	7	▮	NaN	NaN	NaN	NaN	NaN	▮	E	I	D
clarity	category	8	▮	NaN	NaN	NaN	NaN	NaN	▮	SI2	VS2	SI2
depth	float64	184	▮	43.00	61.75	61.80	79.00	1.43	▮	61.5	59.3	62.2
table	float64	127	▮	43.00	57.46	57.00	95.00	2.23	▮	55.0	59.0	55.0
price	int64	11,602	▮	326.00	3,932.80	2,401.00	18,823.00	3,989.44	▮	326	2300	2757
x	float64	554	▮	0.00	5.73	5.70	10.74	1.12	▮	3.95	5.89	5.83
y	float64	552	▮	0.00	5.73	5.71	58.90	1.14	▮	3.98	5.81	5.87
z	float64	375	▮	0.00	3.54	3.53	31.80	0.71	▮	2.43	3.47	3.64

Contents

Pandas - DataFrame의 각 columns 분석을 위한 유틸리티 함수

1. 핵심 요약

2. 사용 예시

2.1. 유틸리티 함수

2.2. titanic

2.3. iris

2.4. diamonds

99. Reference

Related Content

Contents

Pandas - DataFrame의 각 columns 분석을 위한 유틸리티 함수

1. 핵심 요약

2. 사용 예시

2.1. 유틸리티 함수

2.2. titanic

2.3. iris

2.4. diamonds

99. Reference

Related Content

Pandas - value_counts(), 고유값의 빈도와 비율을 같이 표시하기

Pandas 그룹내에서 순서대로 인덱스 부여 - cumcount()

Pandas Dataframe 구조 변경 - unstack()

Pandas DataFrame 함수 매핑 - apply()

Pandas DataFrame 그룹 연산 - groupby()