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

import seaborn as sns
sns.set_style('whitegrid')
import warnings
warnings.filterwarnings('ignore')


from scipy import stats
from sklearn import linear_model
from sklearn import neighbors
from sklearn.metrics import mean_squared_error
from sklearn import preprocessing
from sklearn.model_selection import train_test_split

from math import log

data=pd.read_csv('kc_house_data.csv',parse_dates=['date'])

data.describe()

data.price.plot(kind='hist',log=True)

data[data['price']>4e6] # 11 houses greated than 4M

plt.figure(figsize=(10,8))
data.bedrooms.plot(),data.bathrooms.plot(),data.waterfront.plot()
plt.legend()

data.sqft_living.plot()

fig=plt.figure(figsize=(6,10))

ax1=plt.subplot(331)
ax2=plt.subplot(332)
ax3=plt.subplot(333)
#ax4=plt.subplot(334)
#ax5=plt.subplot(335)
#ax6=plt.subplot(336)
#ax7=plt.subplot(337)


data.boxplot(column='price',ax=ax1)
data.boxplot(column='bedrooms',ax=ax2)
data.boxplot(column='bathrooms',ax=ax3)


plt.suptitle('')
plt.tight_layout()

data[data['bedrooms']>10]

outliers=data.quantile(0.90)
x=data[(data['price']<outliers['price'])]
x=x[(x['bedrooms']< outliers['bedrooms'])]
x=x[(x['bathrooms']< outliers['bathrooms'])]
x=x[(x['sqft_living']< outliers['sqft_living'])]
x.shape

data.shape

x_zipcode=pd.get_dummies(x['zipcode'],drop_first=True)

x=pd.concat([x,x_zipcode],axis=1)

x['built_ago']=2017-x['yr_built']
x['have_basement']=np.where(x['sqft_living']>0,1,0)
x['renovated']=np.where(x['yr_renovated']>0,1,0)
x['weighted_bath']=x['bathrooms'] **2
x['weighted_livingspace']=x['sqft_living']**2
x['diff_living']=x['sqft_living']-x['sqft_living15']
x['bed_bath_ratio']=(x['bedrooms']+1)/(x['bathrooms']+1)

y=x.price
x=x.drop(['id','date','zipcode','lat','long','price','yr_built','sqft_basement','bathrooms'],axis=1) # all of them id date zipcode lat long price yr_renovated yr_built sqft_basement bathrooms grade

# Linear Regression
x_train,x_test,y_train,y_test=train_test_split(x,y,train_size=0.8,random_state=42)
x_train.shape,y_train.shape,x_test.shape,y_test.shape

reg=linear_model.LinearRegression()
regmodel=reg.fit(x_train,y_train)
y_predtest=reg.predict(x_test)

RMS=mean_squared_error(y_test,y_predtest) ** 0.5
RMS

from sklearn.linear_model import Lasso
ls=Lasso()
l=ls.fit(x_train,y_train)
y_ls_predtest=ls.predict(x_test)
ls_rmse=mean_squared_error(y_test,y_ls_predtest) ** 0.5
ls_rmse