import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
np.random.seed(1) 
temp= np.array([-10.2, -5.2, 0.1, 10.1, 12.2, 14.7,25.4, 26.8, 28.9, 35.1, 32.2, 34.6])
ϵ1= np.random.normal(size=12,scale=5)
icecream= 20 + temp * 2 + ϵ1 

np.random.seed(2) 
ϵ2= np.random.normal(size=12,scale=5) 
disease = 30+ temp* 0.5 + ϵ2

plt.plot(temp,icecream,'.')

[<matplotlib.lines.Line2D at 0x1c29099ffd0>]
plt.plot(temp,disease,'.')

[<matplotlib.lines.Line2D at 0x1c291128cd0>]
plt.plot(icecream,disease,'.') # 아이스크림 & 소아마비

[<matplotlib.lines.Line2D at 0x1c2911a0940>]
plt.plot(icecream[:6],disease[:6],'.')
# 비슷한 온도로 관찰 -> 선형관계 약해짐

[<matplotlib.lines.Line2D at 0x1c29120c5e0>]
df=pd.read_csv('https://raw.githubusercontent.com/guebin/2021DV/master/_notebooks/extremum.csv')

df.columns

Index(['지점번호', '지점명', '일시', '평균기온(℃)', '최고기온(℃)', '최고기온시각', '최저기온(℃)',
       '최저기온시각일교차', 'Unnamed: 8'],
      dtype='object')
pd.Series(df.columns)

0          지점번호
1           지점명
2            일시
3       평균기온(℃)
4       최고기온(℃)
5        최고기온시각
6       최저기온(℃)
7     최저기온시각일교차
8    Unnamed: 8
dtype: object
temp=np.array(df.iloc[:,3]) # 평균기온 열
len(temp) #평균기온에 해당되는 행 656개

656
np.random.seed(1)
ϵ1=np.random.normal(size=656, scale=10) 
icecream=temp*2 + 30 + ϵ1 

np.random.seed(2) 
ϵ2=np.random.normal(size=656,scale=1)
disease=temp*0.5 + 40 +ϵ2

plt.plot(temp,icecream,'.')

[<matplotlib.lines.Line2D at 0x1c292285490>]
plt.plot(temp,disease,'.')

[<matplotlib.lines.Line2D at 0x1c2922ec310>]
plt.plot(icecream,disease,'.')

[<matplotlib.lines.Line2D at 0x1c292342a30>]
np.corrcoef(icecream,disease) #상관계수 -> 인과관계 의미 x

array([[1.        , 0.86298975],
       [0.86298975, 1.        ]])
plt.plot(icecream[temp>25],disease[temp>25], '.') # 비슷한 온도구간 관찰 -> 선형관계 약해짐

[<matplotlib.lines.Line2D at 0x1c29283bb80>]
fig , ((ax1,ax2), (ax3,ax4)) = plt.subplots(2,2) 
ax1.plot(temp,icecream,'.')
ax2.plot(temp,disease,'.')
ax3.plot(icecream,disease,'.')
ax4.plot(icecream,disease,'.')
ax4.plot(icecream[temp>25],disease[temp>25],'.')

[<matplotlib.lines.Line2D at 0x1c292967430>]

온도구간 카테고리화 

df1=pd.DataFrame({'temp':temp, 'icecream':icecream, 'disease':disease})

df1
temp icecream disease
0 -0.5 45.243454 39.333242
1 1.4 26.682436 40.643733
2 2.6 29.918282 39.163804
3 2.0 23.270314 42.640271
4 2.5 43.654076 39.456564
... ... ... ...
651 19.9 78.839992 49.633906
652 20.4 86.554679 48.920443
653 18.3 78.666079 49.882650
654 12.8 52.771364 46.613159
655 6.7 40.736731 44.902513

656 rows × 3 columns

df1.temp

0      -0.5
1       1.4
2       2.6
3       2.0
4       2.5
       ... 
651    19.9
652    20.4
653    18.3
654    12.8
655     6.7
Name: temp, Length: 656, dtype: float64
df1.temp.hist() # = plt.hist(df1.temp)

<AxesSubplot:>
def f(x): 
    if x<0: y='group0'
    elif x<10: y='group10'
    elif x<20: y='group20'
    else: y='group30'
    return y 

df1['temp2']=list(map(f,df1.temp))

df1
temp icecream disease temp2
0 -0.5 45.243454 39.333242 group0
1 1.4 26.682436 40.643733 group10
2 2.6 29.918282 39.163804 group10
3 2.0 23.270314 42.640271 group10
4 2.5 43.654076 39.456564 group10
... ... ... ... ...
651 19.9 78.839992 49.633906 group20
652 20.4 86.554679 48.920443 group30
653 18.3 78.666079 49.882650 group20
654 12.8 52.771364 46.613159 group20
655 6.7 40.736731 44.902513 group10

656 rows × 4 columns

from plotnine import * 
ggplot(df1)+geom_point(aes(x='icecream',y='disease',color='temp2'))

<ggplot: (120959691903)>
ggplot(df1)+geom_point(aes(x='icecream',y='disease',color='temp2'))\
+geom_smooth(aes(x='icecream',y='disease',colour='temp2'),size=1,linetype='dashed')

C:\Users\ehfus\Anaconda3\envs\dv2021\lib\site-packages\plotnine\stats\smoothers.py:310: PlotnineWarning: Confidence intervals are not yet implementedfor lowess smoothings.

<ggplot: (120949679559)>
ggplot(df1,aes(x='icecream',y='disease'))+geom_point(aes(color='temp2'))\
+geom_smooth(aes(color='temp2'),size=1,linetype='dashed')

C:\Users\ehfus\Anaconda3\envs\dv2021\lib\site-packages\plotnine\stats\smoothers.py:310: PlotnineWarning: Confidence intervals are not yet implementedfor lowess smoothings.

<ggplot: (120959836549)>

만약 아이스크림과 질병이 연관있는 경우라면

np.random.seed(1)
ϵ1=np.random.normal(size=656, scale=10) 
icecream=temp*2 + 30 + ϵ1 
np.random.seed(2) 
ϵ2=np.random.normal(size=656,scale=1)
disease= 30+ temp*0.0 + icecream*0.15 +ϵ2*2 # 식 자체 정의를 달리해야 함

df2=pd.DataFrame({'temp':temp,'icecream':icecream,'disease':disease})
df2['temp2']=list(map(f,df2.temp)) # map을 이용하여 기존df에 자료 추가해주려면 list로 !!

ggplot(df2)+geom_point(aes(x='icecream',y='disease',colour='temp2'))\
+geom_smooth(aes(x='icecream',y='disease',colour='temp2'),size=2,linetype='dashed')

C:\Users\ehfus\Anaconda3\envs\dv2021\lib\site-packages\plotnine\stats\smoothers.py:310: PlotnineWarning: Confidence intervals are not yet implementedfor lowess smoothings.

<ggplot: (120959800808)>
df1.corr()
temp icecream disease
temp 1.000000 0.884366 0.975609
icecream 0.884366 1.000000 0.862990
disease 0.975609 0.862990 1.000000 
df2.corr()
temp icecream disease
temp 1.000000 0.884366 0.725505
icecream 0.884366 1.000000 0.830539
disease 0.725505 0.830539 1.000000