commute - salary 分布の重なり

import numpy as np 
import pandas as pd 
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.preprocessing import LabelEncoder

root = '../data/input/'

train_df = pd.read_csv(root + 'train_data.csv')
test_df = pd.read_csv(root + 'test_data.csv')

def preprocess(df):
    df["area"] = LabelEncoder().fit_transform(df["area"])
    return df
train_df = preprocess(train_df)
test_df = preprocess(test_df)

plt.figure(figsize=(10,10))

plt.scatter(train_df["commute"], train_df["salary"])
fs = 22
roundnum = 4

plt.xlabel('commute',fontsize=fs)
plt.ylabel('salary',fontsize=fs)
corr = round(train_df.corr()["commute"]["salary"], roundnum)
plt.title("commute - salary\nR: " + str(corr), fontsize=fs)
plt.tick_params(labelsize=fs)

train_p_1 = train_df[train_df["partner"]==1]
train_p_0 = train_df[train_df["partner"]==0]



plt.figure()
fig, ax = plt.subplots(2,1,figsize=(20,10))

fs = 22

plt.subplot(1,2,1)
plt.scatter(train_df["commute"], train_df["salary"])

plt.xlabel('commute',fontsize=fs)
plt.ylabel('salary',fontsize=fs)
corr = round(train_df.corr()["commute"]["salary"],roundnum)
plt.title("commute - salary\nR: " + str(corr), fontsize=fs)
plt.tick_params(labelsize=fs)






plt.subplot(1,2,2)
feature = "area"

df1 = train_p_1[(train_p_1[feature]==9) |(train_p_1[feature]==26)]
plt.scatter(df1["commute"], df1["salary"], label='1, married & working in tokyo or osaka')

df2 = train_p_1[(train_p_1[feature]!=9) &(train_p_1[feature]!=26)]
plt.scatter(df2["commute"], df2["salary"],label="2, married & working except in tokyo or osaka")

df3 = train_p_0[(train_p_0[feature]==9) |(train_p_0[feature]==26)]
plt.scatter(df3["commute"], df3["salary"],label="3, unmarried & working in tokyo or osaka")

df4 = train_p_0[(train_p_0[feature]!=9) &(train_p_0[feature]!=26)]
plt.scatter(df4["commute"], df4["salary"],label="4, unmarried & working iexcept in tokyo or osaka")

plt.legend(fontsize=15,loc='lower right')
plt.xlabel('commute',fontsize=fs)
plt.ylabel('salary',fontsize=fs)
plt.title("4 types of 'commute - salary'", fontsize=fs)
plt.tick_params(labelsize=fs)

plt.figure()
fig, ax = plt.subplots(2,2,figsize=(22,22))
cmap = plt.get_cmap("tab10") 

plt.subplot(2,2,1)
corr = round(df1.corr()["commute"]["salary"],roundnum)
plt.scatter(df1["commute"], df1["salary"],color=cmap(0), label=corr)
plt.xlabel('commute',fontsize=fs)
plt.ylabel('salary',fontsize=fs)
plt.title('1, married & working in tokyo or osaka\nR: '+str(corr),fontsize = fs)
plt.tick_params(labelsize=fs)


plt.subplot(2,2,2)
plt.scatter(df2["commute"], df2["salary"],color=cmap(1))
corr = round(df2.corr()["commute"]["salary"],roundnum)
plt.xlabel('commute',fontsize=fs)
plt.ylabel('salary',fontsize=fs)
plt.title("2, married & working except in tokyo or osaka\nR: "+str(corr),fontsize=fs)
plt.tick_params(labelsize=fs)


plt.subplot(2,2,3)
plt.scatter(df3["commute"], df3["salary"],color=cmap(2))
corr = round(df3.corr()["commute"]["salary"],roundnum)
plt.xlabel('commute',fontsize=fs)
plt.ylabel('salary',fontsize=fs)
plt.title("3, unmarried & working in tokyo or osaka\nR: "+str(corr),fontsize=fs)
plt.tick_params(labelsize=fs)


plt.subplot(2,2,4)
plt.scatter(df4["commute"], df4["salary"],color=cmap(3))
corr = round(df4.corr()["commute"]["salary"],roundnum)
plt.xlabel('commute',fontsize=fs)
plt.ylabel('salary',fontsize=fs)
plt.title("4, unmarried & working iexcept in tokyo or osaka\nR: "+str(corr),fontsize=fs)
plt.tick_params(labelsize=fs)
plt.subplots_adjust(hspace=0.3)