Linear_Reg

# -*- coding: utf-8 -*-
"""
Created on Fri Jul 12 10:08:27 2019

@author: Ritvik Sapra
"""

'''
We first try to visualise data. We try to understand in terms of the three states
about their the profit and expenditure spent by them.
We then use Machine Lerning model for prediction and finally find out the accuracy of model.
'''

#%%
''' We first load the data set from the PC.'''
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sb
dataset=pd.read_csv(r"C:\Users\nitik\Desktop\Internal_ML_Project_Final\50_Startups.csv")
dataset=dataset.fillna(0)
dataset
#%%
''' We now make a pie chart of R&D Spend, Administration, Marketing Spend and Profit of 
the three states. '''
data=dataset.groupby(['State']).sum ()
print("The table for sum() grouped by 'State' is given below:")
print(data)

A=data.iloc[:,-1]
B=data.iloc[:,0]
C=data.iloc[:,1]
D=data.iloc[:,2]

l=['California','Florida','New York']
c=['r','b','g']

plt.pie(B,labels=l,colors=c,autopct='%1.01f%%',explode=(0,0,0.1),startangle=90)
plt.xlabel('x axis---->')
plt.ylabel('y axis----->')
plt.title('R&D Spend')
plt.legend()
plt.show()

plt.pie(C,labels=l,colors=c,autopct='%1.01f%%',explode=(0,0,0.1),startangle=90)
plt.xlabel('x axis---->')
plt.ylabel('y axis----->')
plt.title('Administration')
plt.legend()
plt.show()

plt.pie(D,labels=l,colors=c,autopct='%1.01f%%',explode=(0,0.1,0),startangle=90)
plt.xlabel('x axis---->')
plt.ylabel('y axis----->')
plt.title('Marketing')
plt.legend()
plt.show()

plt.pie(A,labels=l,colors=c,autopct='%1.01f%%',explode=(0,0,0.1),startangle=90)
plt.xlabel('x axis---->')
plt.ylabel('y axis----->')
plt.title('Profit')
plt.legend()
plt.show()
#%%
''' We have to predict Profit. Thus, we plot between other columns and Profit
to understand their relation. '''
#between R&D and Profit
plt.plot(B,A)
plt.title('R&D Spend vs. Profit')
plt.xlabel('R&D Spend')
plt.ylabel('Profit')
plt.show()

#between Administration and Profit
plt.plot(C,A)
plt.title('Administration vs. Profit')
plt.xlabel('Administration')
plt.ylabel('Profit')
plt.show()

#between Marketing and Profit
plt.plot(D,A)
plt.title('Marketting vs. Profit')
plt.xlabel('Marketing')
plt.ylabel('Profit')
plt.show()
#%%
''' We now do the pair plot to understand properly relation between all the data. '''
sb.pairplot(dataset)
#%%
''' We now know that profit is dependent on R&D Spend and Marketing. To understand it better we 
see the correlation table '''
print(dataset.corr())
#%%
''' Since now we know our 'x' set and 'y' set, we assign the data sets to the variables. '''
x=dataset.iloc[:,[0,2]].values
y=dataset.iloc[:,-1].values
print(x)
print(y)
#%%
''' Now we have the data sets ready. We now plot a relation between the the two sets. '''
#plot
plt.plot(x,y)
#%%
#reg plot between R&D Spend and Profit
sb.regplot(dataset['R&D Spend'],y)
#%%
#reg plot between Marketing Spend and Profit
sb.regplot(dataset['Marketing Spend'],y)
#%%
''' We have properly visualised the data and found out the reltion. Now we implement
the ML model, i.e., Linear Regresion (we found it most effective amongst other models) '''

#We train, test and split the data.
from sklearn.model_selection import train_test_split
x_train,x_test,y_train,y_test=train_test_split(x,y,test_size=0.25,random_state=21)
x_train, y_train
#%%
from sklearn.linear_model import LinearRegression
lr=LinearRegression()
lr.fit(x_train,y_train)
pred1=lr.predict(x_test)
pred1
#%%
from sklearn.metrics import r2_score
acc1=r2_score(y_test,pred1)
acc1
#%%