# data analysis workshop 1, 13.4.2021

# example, car sales

import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from scipy import stats

def percentage_difference(row):
    selling = row['Selling_Price']
    road = row['Present_Price']
    result = 1 - round(selling / road, 2)
    return result


cars = pd.read_csv('car data.csv')

cars = cars.drop('Car_Name', axis=1)

cars['Price_Difference'] = cars.apply(percentage_difference, axis=1)

column_names = ['Year', 'Selling_Price', 'Present_Price', 'Price_Difference', 'Kms_Driven', 'Fuel_Type', 'Seller_Type', 'Transmission', 'Owner']
cars = cars.reindex(columns=column_names)

cars = cars.drop('Owner', axis=1)

automatic_diesels = cars[cars['Fuel_Type'] == 'Diesel']

# the amount of extra costs (price difference) is affected by car year and kms driven
correlations = cars.corr()

plt.clf()
sns.pairplot(cars)
plt.figure()


# Automatic transmission cars tend to be more expensive
plt.clf()
sns.pairplot(cars, hue='Transmission')
plt.figure()

# Also Diesel fuel type seem to be more valuable in used cars
plt.clf()
sns.pairplot(cars, hue='Fuel_Type')
plt.figure()

plt.clf()
sns.boxplot(x='Transmission', y='Present_Price', data=cars, hue='Fuel_Type')
plt.figure()