[機械学習講座]4データ解析用Python(5)

データFrame Booleanとしてデータを選択

import pandas as pd

# data 출처: https://www.kaggle.com/hesh97/titanicdataset-traincsv/data
train_data = pd.read_csv('../train.csv')

train_data.head()

class_ = train_data['Pclass'] == 1
age_ = (train_data['Age'] >= 30) & (train_data['Age'] < 40)

train_data[class_ & age_]

列の追加または削除

train_data['Age_double'] = train_data['Age'] * 2
train_data.head()

train_data['Age_tripple'] = train_data['Age_double'] + train_data['Age']
train_data.head()

train_data.insert(3, 'Fare10', train_data['Fare'] / 10)
train_data.head()

train_data.drop('Age_tripple', axis=1)
train_data.head()

train_data.drop('Age_double', axis=1)
train_data.head()

train_data.drop(['Age_double', 'Age_tripple'], axis=1, inplace=True)

DataFrame列間の相関の計算

import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline 

# data 출처: https://www.kaggle.com/hesh97/titanicdataset-traincsv/data
train_data = pd.read_csv('./train.csv')
train_data.head()

train_data.corr()

plt.matshow(train_data.corr())