import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn import datasets
from sklearn.model_selection import train_test_split
from sklearn.datasets import load_boston
import seaborn as sns
from sklearn import preprocessing
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import r2_score
import matplotlib.pyplot as plt
from sklearn import preprocessing

boston = load_boston()

print(boston)

print(boston.keys())

print(boston.data.shape)

print(boston.feature_names)

print(boston.DESCR)

boston_df = pd.DataFrame(boston.data)
boston_df.columns = boston.feature_names
print(boston_df.head())

print(boston.target.shape)

boston_df['PRICE'] = boston.target
print(boston_df.head())

print(boston_df.describe)

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn import datasets
from sklearn.model_selection import train_test_split
from sklearn.datasets import load_boston
import seaborn as sns
from sklearn import preprocessing
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import r2_score
import matplotlib.pyplot as plt
from sklearn import preprocessing
from mpl_toolkits.mplot3d import Axes3D

boston = load_boston()
boston_df = pd.DataFrame(boston.data)
boston_df.columns = boston.feature_names
boston_df['PRICE'] = boston.target

# Setting up the variables
# We want to predict house price based on crime rate and proximity to industry
feature_cols = ['CRIM', 'INDUS']
X = pd.DataFrame(boston_df, columns= feature_cols)
y = pd.DataFrame(boston_df, columns = ['PRICE'])

ya = np.array(y).reshape((-1,))
x_crime = np.array(X['CRIM'])
x_indus = np.array(X['INDUS'])

# Now let's do a multiple regression using both quantitative predictors
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state = 0 )
linModel = LinearRegression()
linModel.fit(X_train, y_train)
print(" ")
print('MULTIPLE LINEAR REGRESSION - predict house price based on crime and proximity to industrial areas')
beta_0 = linModel.intercept_
print("Beta_0 (y-intercept) for the model is ", beta_0)
# linModel.coef_ returns the beta coefficients for slope (in same order as passed  --  beta_1, beta_2, etc. for multiple regression)
beta_1_1 = linModel.coef_[0,0]
print("Beta_1 (slope) for predictor ", feature_cols[0], " is ", beta_1_1)
beta_1_2 = linModel.coef_[0,1]
print("Beta_1 (slope) for predictor ", feature_cols[1], " is ", beta_1_2)
y_pred = linModel.predict(X_test)
linModel_score = r2_score(y_test, y_pred)
print ("R^2 score for the model is ", linModel_score)

# do 3d plot of observations
fig = plt.figure(figsize=(12,8))
ax = fig.gca(projection='3d')
ax.scatter3D(x_crime, x_indus, ya, c=ya, cmap='Reds', s=80)
ax.set_xlabel('Crime', fontsize=20)
ax.set_ylabel('Industrial Proximity', fontsize=20)
ax.set_zlabel('House Price', fontsize=20)
plt.title('Actual Observations', fontsize=20)
ax.legend()
plt.show()

#compute estimated tips from the linear model
def f(x1, x2, b0, b1, b2):
    return b0 + b1 * x1 + b2 * x2

est_price = f(x_crime, x_indus, beta_0, beta_1_1, beta_1_2)

#compare estimated tips from the linear model and actual house price
fig = plt.figure(figsize=(12,8))
plt.plot(ya)
plt.plot(est_price)
plt.show()

# compute a surface for the estimated housing price
xx1, xx2 = np.meshgrid(x_crime, x_indus)
Z = f(xx1, xx2, beta_0, beta_1_1, beta_1_2)
# do 3d plot
fig = plt.figure(figsize=(12,8))
ax = fig.gca(projection='3d')
ax.plot_wireframe(xx1, xx2, Z, rstride=5, cstride=6,  alpha=0.2) #cmap='cubehelix', #
ax.scatter3D(x_crime, x_indus, ya, color = 'green', s=80) #c=ya,
ax.set_xlabel('Crime', fontsize=20)
ax.set_ylabel('Industrial Proximity', fontsize=20)
ax.set_zlabel('House Price', fontsize=20)
ax.set_zlim(0, ya.max())
plt.title('Linear Model to Estimate/Predict House Price', fontsize=20)
ax.legend()
plt.show()

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn import datasets
from sklearn.model_selection import train_test_split
from sklearn.datasets import load_boston
import seaborn as sns
from sklearn import preprocessing
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import r2_score
import matplotlib.pyplot as plt
from sklearn import preprocessing
from mpl_toolkits.mplot3d import Axes3D

boston = load_boston()
boston_df = pd.DataFrame(boston.data)
boston_df.columns = boston.feature_names
boston_df['PRICE'] = boston.target

# Setting up the variables
# We want to predict house price based on crime rate and proximity to industry
feature_cols = np.array(['CRIM', 'INDUS'])
X_pdf = pd.DataFrame(boston_df, columns= feature_cols)
y = pd.DataFrame(boston_df, columns = ['PRICE'])

# Linear regression on the two predictors individually
for i in range(len(feature_cols)):
    X = X_pdf[feature_cols[i]]
    X = X.values.reshape(-1, 1)  # a formatting trick required by the built in regression functions in sklearn
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state = 0 )
    linModel = LinearRegression()
    linModel.fit(X_train, y_train)
    print(" ")
    print('SIMPLE LINEAR REGRESSION FOR PREDICTORS INDIVIDUALLY - predict house price from crime rate and proximity to industrial areas')
    beta_0 = linModel.intercept_
    print("Beta_0 (y-intercept) for model with predictor ", feature_cols[i], " is ", beta_0)
    beta_1 = linModel.coef_[0]
    print("Beta_1 (slope) for model with predictor ", feature_cols[i], " is ", beta_1)
    y_pred = linModel.predict(X_test)
    linModel_score = r2_score(y_test, y_pred)
    print ("R^2 score for model with predictor ", feature_cols[i], " is ", linModel_score)
    plt.subplot(1, 2, i+1)
    plt.scatter(X, y)
    plt.plot(X, beta_0 + beta_1*X)
    plt.ylim(0, )
    plt.title('Simple linear regression for predictor ' + feature_cols[i], fontsize=18)
    plt.text(0,0, "R^2 score = " + str(round(linModel_score,4)), fontsize=18)
    if i==0: plt.ylabel('House Price in Thousands of Dollars', fontsize=18)
plt.show()