Credit Applicants, Visualized

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

df = pd.read_csv('clean_dataset.csv')
df.head()

#1
sns.catplot(data=df, y='Debt', x='Citizen', kind= 'box', hue= 'Citizen')
plt.xticks(rotation=20)
plt.show()

#On average, people who are born in the country tend
#have more debt than those who are in the coutnry by other means, and temporary

#2
#age and debt corr
plt.figure(figsize=(8,4))
sns.jointplot(data=df, x='Age', y='Debt', kind='reg')
plt.show()

#age and income corr
plt.figure(figsize=(8,4))
sns.jointplot(data=df, x='Age', y='Income', kind='reg')
plt.show()

#debt and income corr
plt.figure(figsize=(8,4))
sns.jointplot(data=df, x='Debt', y='Income', kind='reg')
plt.show()

#We found that the strongesrt correlation is between Age and debr dramatically more
#than the correlations between debt and income, and lastly by age and income

#3

Industry_count= df['Industry'].value_counts()
Industry_count

Industry_pie= df['Industry'].value_counts().index.tolist()
Industry_pie

plt.figure(figsize=(6,6))
plt.pie(Industry_count, labels=Industry_pie, autopct= '%1.2f%%')
plt.show()

#By looking at the pie chart above, we foudn that the only industry group compromising at leats 20%
#of the overall customer population is Energy.

#4
plt.figure(figsize=(6,5))
sns.scatterplot(data=df, y='Income', x='Debt', hue='Approved', marker='x', s=50)
plt.show()

#Based on the scatter plot above on Debt and Income, we found that a person will usually have
#a higher chance of getting approved for credit if they have a higher income and less debt

#5
plt.figure(figsize=(10,5))
sns.kdeplot(data=df, x='Income', hue='Ethnicity', fill=True,
            alpha=0.2, linewidth=1, linestyle='-', common_norm=False)
plt.show()

#Based on the kde plot above, we can see that the ethnicity with the smallest income gap
#is Latino and majority of their distrubution having about 0 income compared to the other ethnicities.

#6
sns.catplot(data=df, y='Debt', x='Married', kind= 'box', hue= 'Married')
plt.show()

#Based on the box plot above we found that on average, married people are more
#likely to have debt. Even with single people having more debt outliers

#7
#plt.figure(figsize=(5,3))
#sns.scatterplot(data=df, x='YearsEmployed', y='Debt')
#plt.show()

sns.scatterplot(data=df, x='YearsEmployed', y='Debt', color='skyblue')
sns.histplot(data=df, x='YearsEmployed', y='Debt', color ='lightblue', pthresh=0.2)
sns.kdeplot(data=df, x='YearsEmployed', y='Debt', color='lightgreen')
plt.show()

#Based off the stacked plot above (scatter, hist, kde), we foudn that people
#who have worked longer tend to have less debt than than those who just started working

#8
sns.catplot(data=df, x= 'Ethnicity', y='Debt', kind= 'box',
            hue= 'Married')
plt.show()

#Based off the box plots above, by looking only at the married couples by ethnicity
#white couples tend to have the lowest debt on average

#9
#define Incomes
attributes= ['Salary', 'Side Hustles', 'Stocks',]
values_chad= [4, 5, 10]
values_chad= values_chad +[4]
values_chad

values_jess= [4, 0, 8]
values_jess= values_jess + [4]
values_jess

#define angles
angles = np.linspace(0, 2*np.pi, 3, endpoint=False).tolist()
angles = angles+ [0]
angles

#initialize the radar chart
plt.figure(figsize=(5,5))
ax= plt.subplot(polar=True)
#add attributes/features
plt.xticks(angles[:-1], attributes)
ax.set_rlabel_position(4)
plt.yticks([2,4,6,8,10], color="blue", alpha=0.8, size=7)
plt.ylim(0,10)
#add Chad info
ax.plot(angles, values_chad, label='Chad',color='blue',linewidth=2, linestyle='--')
ax.fill(angles, values_chad, alpha=0.4, facecolor='pink')
#add Jess info
ax.plot(angles, values_jess, label='Jess',linewidth=2, color='orange', linestyle='--')
ax.fill(angles, values_jess, alpha=0.2, facecolor='lightgreen')
plt.legend(loc='upper left')
plt.show()

#the visualization above shows the comparison between Chad and jess's sources of income after being normalized.
#they show that Chad over all makes more money

#10
def func_(x):
  outcome= 2000 +x**2
  return outcome

x = np.linspace(0,12, 100) + 6
y = func_(x)

plt.figure(figsize=(4, 4))
plt.plot(x, y, color='indigo', linestyle='--', linewidth=3)
plt.axvline(x=6, color= 'red', linestyle= '--', linewidth=1.3)
plt.grid()
plt.show()

#the graph above shows the savings increasing after month 6

#11
def func1_(x):
  outcome= x**2 + 2 * x + 12
  return outcome

def func2_(x):
  outcome= 3 * x + 12
  return outcome

x = np.linspace(1,10,10)
sales = func1_(x)
profit = func2_(x)

plt.figure(figsize=(5,3))
sns.pointplot( x= sales, y= profit, linestyle='--',
              marker='X', markersize= '15', color='hotpink', mfc='steelblue',
              mec='black')
plt.grid(axis= 'y', linewidth=0.5, linestyle='--', color= 'blue')
plt.show()

#Based on the graph above, we can summarize the relationship between sales and profit to have a
#positive relationship, and they both increase together

#12
import networkx as nx # for networks

# initialize the graph/ network
G1 = nx.DiGraph() #directed graph
#add nodes
G1.add_node(1)
G1.add_node(2)
G1.add_node(3)
G1.add_node(4)

G1.add_edge(1, 2)
G1.add_edge(1, 3)
G1.add_edge(1, 4)
G1.add_edge(2, 3)
G1.add_edge(2, 4)

plt.figure(figsize=(5,5))
pos = nx.spring_layout(G1) #define the position in whch our nodes can be visualized, spring layout is the most common used
nx.draw_networkx_nodes(G1, pos, node_size=500, node_color='#fb8500')
nx.draw_networkx_edges(G1, pos, edge_color='#8338ec', width=2)
nx.draw_networkx_labels(G1, pos, font_color='#ede0d4')
plt.show()

#the visual above shows the comminucation between managers (1 and 2), and employees (3 and 4).
#managers can communicate with one another and employees, but employees can only communicate with managers

jupyter nbconvert --to html bana310_HW2and3.ipynb