import pandas as pd
import matplotlib.pyplot as plt
import numpy as np


donnees = pd.read_csv('http://louismerlin.fr/Enseignement/2223/TP/tp2_data.csv', sep =';')

# print(donnees.head())

table1=donnees[['Année', 'taux fertilité j. femmes', 'femmes scol. sec.']]
table1=table1.rename(columns={'taux fertilité j. femmes': 'TF', 'femmes scol. sec.':'FSS'})

# print(table1.mean())
# print(table1.std())

table1=table1.dropna()
X = table1['FSS']
Y = table1['TF']

## Dessin du nuage de points.


# plt.grid()
#plt.plot(X,Y,'k+')
# plt.show()

## Droite de régression.

# a = X.cov(Y) / ((X.std())**2)
# b = Y.mean()-a*X.mean()

# t=np.arange(51,67,0.01)

# plt.grid()
# plt.plot(X,Y,'k+')
# plt.plot(t,a*t+b)
# plt.show()


## Exercice 1


# table2=donnees[['Année', 'conso élec per capita', 'esperance vie']]
# table2=table2.dropna()

# A = table2['conso élec per capita']
# B = table2['esperance vie']

# plt.plot(A,B,'k+')
# plt.show()

# a = A.cov(B) / ((A.std())**2)
# b = B.mean()-a*A.mean()

# t=np.arange(2100,3300,1)
# plt.plot(A,B,'k+')
# plt.plot(t,a*t+b)
# plt.show()



## Regression avec transformation

# data2=pd.read_csv('https://louismerlin.fr/Enseignement/2223/TP/tp2_nor.csv', sep=';')
# data2=data2.dropna()


# X=data2['PIB per capita']
# Y=data2['Pop urbaine %']

# plt.grid()
# plt.plot(X,Y, '.') # nuage de points
# plt.show()

# Xlog = np.log(X)
# plt.plot(Xlog,Y,'.')
# plt.show()

# rho = Xlog.cov(Y) /(Xlog.std()*Y.std())
# print(rho)

# a = Xlog.cov(Y)/((Xlog.std())**2)
# b = Y.mean()-a*Xlog.mean()

# t=np.arange(100,100000,10)
# plt.plot(X,Y,'.')
# plt.plot(t,a*(np.log(t))+b)
# plt.show()