Data Exploration and Visualization (PY)

#we will use pandas to read_csv function and create a data frame called d1:
import pandas as pd
d1 = pd.read_csv("USA_data.csv")
#to make sure we did this correctly we will view the first 3 elements of d1:
d1.head(3)

#next we will perfom a summary of our dataframe using describe:
d1.describe()

#We will now find the mean of just Real GDP 
#notice how the name of the variable goes in brackets after the dataframe d1:
(d1["Real GDP"]).mean()

#a similar logic applies to find the standard deviation of the savings:
(d1["Savings"]).std()

#now we will find the median for the taxes series:
(d1["Taxes"]).median()

#we will call the new variable we want to add Savings Rate 
#we can see from the syntax this is just a vectorial operation:
d1["Savings rate"] = d1["Savings"]/d1["Real GDP"]
#checking if this new variable was added:
d1.head(3)

#to finish we will find the mean savings proportion for this time period:
(d1["Savings rate"].mean())*100

#loading matplotlib
import matplotlib.pyplot as plt

#to shorten our sintax we will create x and y, with Taxes and Real GDP respectively:
x = d1["Taxes"]
y = d1["Real GDP"]

#now for the basic plot:
plt.plot(x, y, 'o', color='midnightblue')
#Note how we specified the "o" which stipulates dots on the plot. 
#Additionally, the color name was added, which is one of my favorites for plots

#Since we want to export this plot to GitHub, we must save the plot as a png file:
plt.savefig('pyplot1.png', dpi=300, bbox_inches='tight')

#now we can add show() to see the plot:
plt.show()

#this is our base plot: 
plt.plot(x, y, 'o', color='midnightblue')

#we import numpy:
import numpy as np

#we use the polyfit funtion, with the independent and dependent
#followed by the order of the polynomial:
m, b = np.polyfit(x, y, 1)

#now we add the trendline by altering our "y" variable:
plt.plot(x, m*x + b, color="firebrick")

#everything else is the same: 
plt.xlabel('Taxes in billions(USD)')
plt.ylabel('Real GDP in billions(USD)')
plt.title('Relationship between taxes and real GDP (1984-2014)')

#saving the plot: 
plt.savefig('pyplot2.png', dpi=300, bbox_inches='tight')
#adding show() to see the plot:
plt.show()

#note that for Seaborn we will use the original names of our variables in the syntax:
import seaborn as sns
sns.set()

#now we can use the Seaborn regplot function:
sns.set_style("dark")
sns.regplot(x="Taxes",
            y="Real GDP", 
            ci=None,data=d1,
            color="midnightblue",line_kws={"color":"firebrick","lw":2})

#extra customization:
plt.title('Relationship between taxes and real GDP (1984-2014)')

#saving the plot: 
plt.savefig('pyplot3.png', dpi=300, bbox_inches='tight')

#adding show() to see the plot:
plt.show()

#importing the style functionality:
import matplotlib.style as style

#we must ask Python to use this FiveThirtyEight style theme for our plot:
style.use('fivethirtyeight')

#this is our base plot: 
plt.plot(x, y, 'o', color='midnightblue')

#we use the polyfit funtion, with the independent and dependent
m, b = np.polyfit(x, y, 1)

#now we add the trendline by altering our "y" variable
plt.plot(x, m*x + b, color="firebrick")

#everything else is the same: 
plt.xlabel('Taxes in billions(USD)')
plt.ylabel('Real GDP in billions(USD)')
plt.title('Relationship between taxes and real GDP (1984-2014)')

#saving our plot:
plt.savefig('pyplot4.png', dpi=300, bbox_inches='tight')

#show the plot:
plt.show()