#!/usr/bin/env python

# coding: utf-8

# In[5]:

import numpy as np

import pandas as pd

import statsmodels.api as sm

import matplotlib.pyplot as plt

from sklearn.linear_model import LinearRegression

import seaborn as sns

sns.set()

# # Loading Raw Data

# In[12]:

raw_data = pd.read_csv('C:\\Users\\Asus\\Desktop\\MachineLearning-Coursera\\LinearRegrssion-CarSales Project\\1.04. Real-life example.csv')

raw_data.head()

# # Preprocessing

# In[13]:

raw_data.describe(include='all')

# In[14]:

data = raw_data.drop(['Model'],axis=1)

# In[15]:

data.describe(include='all')

# ### Dealing with missing values

# In[16]:

data_no_mv = data.dropna(axis=0)

# In[17]:

data_no_mv.describe(include='all')

# # Exploring pdfs

# In[18]:

sns.distplot(data_no_mv['Price'])

# # Dealing with Outliers

# In[19]:

q = data_no_mv['Price'].quantile(0.99)

data_1 = data_no_mv[data_no_mv['Price']<q]

data_1.describe(include='all')

# In[20]:

sns.distplot(data_1['Price'])

# In[21]:

sns.distplot(data_1['Mileage'])

# In[22]:

q = data_1['Mileage'].quantile(0.99)

data_2 = data_1[data_1['Mileage']<q]

# In[23]:

sns.distplot(data_2['Mileage'])

# In[24]:

data_3 = data_2[data_2['EngineV']<6.5]

# In[25]:

sns.distplot(data_3['EngineV'])

# In[26]:

sns.distplot(data_no_mv['Year'])

# In[27]:

q = data_3['Year'].quantile(0.01)

data_4 = data_3[data_3['Year']>q]

# In[28]:

sns.distplot(data_4['Year'])

# In[29]:

data_cleaned = data_4.reset_index(drop=True)

# In[30]:

data_cleaned.describe(include='all')

# # Checking OLS Assumptions

# In[31]:

f, (ax1, ax2, ax3) = plt.subplots(1, 3, sharey=True, figsize=(15,3))

ax1.scatter(data_cleaned['Year'], data_cleaned['Price'])

ax1.set_title('Price and Year')

ax2.scatter(data_cleaned['EngineV'], data_cleaned['Price'])

ax2.set_title('Price and EngineV')

ax3.scatter(data_cleaned['Mileage'], data_cleaned['Price'])

ax3.set_title('Price and Mileage')

plt.show()

# In[32]:

sns.distplot(data_cleaned['Price'])

# ### Relaxing Assumptions

# In[33]:

log_price = np.log(data_cleaned['Price'])

data_cleaned['log_price']=log_price

data_cleaned

# In[34]:

f, (ax1, ax2, ax3) = plt.subplots(1, 3, sharey=True, figsize=(15,3))

ax1.scatter(data_cleaned['Year'], data_cleaned['log_price'])

ax1.set_title('Log Price and Year')

ax2.scatter(data_cleaned['EngineV'], data_cleaned['log_price'])

ax2.set_title('Log Price and EngineV')

ax3.scatter(data_cleaned['Mileage'], data_cleaned['log_price'])

ax3.set_title('Log Price and Mileage')

plt.show()

# In[35]:

data_cleaned = data_cleaned.drop(['Price'],axis=1)

# ## Multicollinearity

# In[36]:

data_cleaned.columns.values

# In[37]:

from statsmodels.stats.outliers_influence import variance_inflation_factor

variables = data_cleaned[['Mileage', 'Year', 'EngineV']]

vif = pd.DataFrame()

vif["VIF"] = [variance_inflation_factor(variables.values, i) for i in range(variables.shape[1])]

vif["features"] = variables.columns

# In[38]:

vif

# In[39]:

data_no_multicollinearity = data_cleaned.drop(['Year'],axis=1)

# # Create Dummy Variables

# In[40]:

data_with_dummies = pd.get_dummies(data_no_multicollinearity, drop_first=True)

# In[41]:

data_with_dummies.head()

# ## Rearrenge

# In[42]:

data_with_dummies.columns.values

# In[43]:

cols = ['log_price','Mileage', 'EngineV', 'Brand_BMW',

'Engine Type_Other', 'Engine Type_Petrol', 'Registration_yes']

# In[44]:

data_preprocessed = data_with_dummies[cols]

data_preprocessed.head()

# # Linear Regression Model

# ## Declare Inputs and Targets

# In[45]:

targets = data_preprocessed['log_price']

inputs = data_preprocessed.drop(['log_price'],axis=1)

# In[46]:

from sklearn.preprocessing import StandardScaler

scaler = StandardScaler()

scaler.fit(inputs)

# In[47]:

inputs_scaled = scaler.transform(inputs)

# ## Train, Test, Split

# In[48]:

from sklearn.model_selection import train_test_split

# In[49]:

x_train, x_test, y_train, y_test= train_test_split(inputs_scaled, targets, test_size=0.2, random_state=365)

# ## Create the Regression

# In[50]:

reg = LinearRegression()

reg.fit(x_train, y_train)

# In[51]:

y_hat = reg.predict(x_train)

# In[53]:

plt.scatter(y_train, y_hat)

plt.xlabel('Targets (y_train)', size=18)

plt.ylabel('Predictions (y_hat)', size=18)

plt.xlim(6,13)

plt.ylim(6,13)

plt.show()

# In[54]:

sns.distplot(y_train - y_hat)

plt.title('Residuals PDF', size=18)

# In[55]:

reg.score(x_train, y_train)

# ## Finding weights and bias

# In[57]:

reg.intercept_

# In[56]:

reg.coef_

# In[58]:

reg_summary = pd.DataFrame(inputs.columns.values, columns=['Features'])

reg_summary['Weights'] = reg.coef_

reg_summary

# In[60]:

data_cleaned['Brand'].unique()

# In[61]:

data_cleaned['Body'].unique()

# In[62]:

data_cleaned['Engine Type'].unique()

# # Testing

# In[63]:

y_hat_test = reg.predict(x_test)

# In[65]:

plt.scatter(y_test, y_hat_test, alpha=0.2)

plt.xlabel('Targets (y_test)', size=18)

plt.ylabel('Predictions (y_hat_test)', size=18)

plt.xlim(6,13)

plt.ylim(6,13)

plt.show()

# In[71]:

df_pf = pd.DataFrame(np.exp(y_hat_test), columns = ['Predictions'])

df_pf.head()

# In[74]:

y_test = y_test.reset_index(drop=True)

y_test.head()

# In[75]:

df_pf = pd.DataFrame(np.exp(y_hat_test), columns = ['Predictions'])

df_pf.head()

# In[80]:

df_pf['Target'] = np.exp(y_test)

df_pf

# In[83]:

df_pf['Prediction'] = np.exp(y_test)

df_pf

# In[84]:

df_pf['Residual'] = df_pf['Target'] - df_pf['Prediction']

# In[85]:

df_pf['Difference%'] = np.absolute(df_pf['Residual']/df_pf['Target']*100)

df_pf

# In[86]:

df_pf.describe()

# In[88]:

pd.options.display.max_rows =999

pd.set_option('display.float_format', lambda x: '%.2f' %x)

df_pf.sort_values(by=['Difference%'])

# In[ ]: