EDA With Python

# Load dataset
billionaires_data_path = r'C:\Users\Francesco\Desktop\PROJECTS\Data Project\Billionaire\data.csv'
billionaires_data = pd.read_csv(billionaires_data_path)

# Display the first 5 rows of the dataframe
billionaires_data_head = billionaires_data.head()
billionaires_data_head

# General information about the dataset
billionaires_data_info = billionaires_data.info()
billionaires_data_info

# List of columns to be dropped
columns_to_drop = ['title', 'date', 'state', 'residenceStateRegion', 'birthYear', 'birthDay', 'birthMonth', 'cpi_change_country', 'latitude_country', 'longitude_country']

# Drop these columns from the dataset
billionaires_data = billionaires_data.drop(columns=columns_to_drop)

# Calculate the number of duplicated rows
billionaires_data.duplicated().sum()

Comment:
There are no duplicated entries in the dataset and we can continue with data cleaning.

# Calculate the number of missing values in each column
billionaires_data.isnull().sum()

# Dropping the 'organization' column due to a high percentage of missing values
billionaires_data = billionaires_data.drop(columns=['organization'])

# Filling missing textual values with "Unknown"
textual_columns = ['country', 'city', 'firstName']
for column in textual_columns:
    billionaires_data[column] = billionaires_data[column].fillna("Unknown")

# Control data types
billionaires_data.dtypes

# Convert a date column to datetime type
billionaires_data['birthDate'] = pd.to_datetime(billionaires_data['birthDate'])

# Converting categorical data
categorical_columns = ['category', 'country', 'city', 'source', 'industries', 
'countryOfCitizenship', 'status', 'gender', 'lastName', 'firstName']
for col in categorical_columns:
    billionaires_data[col] = billionaires_data[col].astype('category')

billionaires_data['gdp_country'] = (
    billionaires_data['gdp_country']
    .replace('[\$,]', '', regex=True)  # Remove $ signs and commas
    .astype(float)  # Convert the result to float)

# This will show only numerical data
billionaires_data.describe()

# Now will show only the other type of data
billionaires_data.describe(include=['category', 'object', 'bool'])

# Histogram of the 'age' column
plt.figure(figsize=(10, 6))
sns.histplot(billionaires_data['age'].dropna(), bins=30, kde=True)
plt.title('Distribution of Age')
plt.xlabel('Age')
plt.ylabel('Frequency')
plt.show()

# Get the count of billionaires by country and select the top 20
country_count = billionaires_data['country'].value_counts().head(20)
# Create the bar plot using Seaborn, which will automatically order the bars
plt.figure(figsize=(12, 8))
sns.barplot(x=country_count.index, y=country_count.values, order=country_count.index)
plt.title('Top 20 Countries by Number of Billionaires')
plt.xlabel('Country')
plt.ylabel('Number of Billionaires')
plt.xticks(rotation=45)
plt.show()

Comment:
This bar chart shows the countries with the highest number of billionaires. It provides a clear view of which countries dominate the billionaire list, is clear that GDP and population can influence this measure, as we will see later.

# Pie chart for the distribution of billionaires by gender
gender_distribution = billionaires_data['gender'].value_counts()
plt.figure(figsize=(8, 8))
plt.pie(gender_distribution, labels=gender_distribution.index, autopct='%1.1f%%')
plt.title('Pie Chart of Billionaires by Gender')
plt.show()

Comment:
The pie chart indicates a significant disparity in the number of male and female billionaires, with males dominating the dataset. This discrepancy highlights gender representation in the context of global wealth distribution, suggesting that males are much more prevalent in the billionaire category than females.

# Extracting the top 20 industries by billionaire count
top_industries_by_billionaires = billionaires_data['industries'].value_counts().head(20)
# Creating the bar chart with the correct order
plt.figure(figsize=(12, 8))
# The 'barh' plot will allow us to place the industry with the most billionaires at the top naturally due to horizontal layout.
top_industries_by_billionaires.sort_values().plot(kind='barh')
plt.title('Top 20 Industries with the Most Billionaires')
plt.xlabel('Number of Billionaires')
plt.ylabel('Industries')
plt.show()

# Boxplot of the 'finalWorth' column
plt.figure(figsize=(10, 6))
sns.boxplot(x=billionaires_data['finalWorth'])
plt.title('Boxplot of Final Worth')
plt.xlabel('Final Worth (in million USD)')
plt.show()

# Function to calculate IQR and identify outliers
def identify_outliers(column_name):
    # Calculate Q1, Q3, and IQR
    Q1 = billionaires_data[column_name].quantile(0.25)
    Q3 = billionaires_data[column_name].quantile(0.75)
    IQR = Q3 - Q1

    # Define the thresholds for outliers
    lower_bound = Q1 - 1.5 * IQR
    upper_bound = Q3 + 1.5 * IQR

    # Identify the outliers in the data
    outliers = billionaires_data[(billionaires_data[column_name] < lower_bound) | (billionaires_data[column_name] > upper_bound)]

    # Summary of outlier detection
    outlier_summary = {
        'Q1': Q1,
        'Q3': Q3,
        'IQR': IQR,
        'Lower Bound': lower_bound,
        'Upper Bound': upper_bound,
        'Number of Outliers': outliers.shape[0]
    }

    return outlier_summary

# Analyzing 'age' column for outliers
age_outlier_summary = identify_outliers('age')

# Plotting boxplot for 'age' column
plt.figure(figsize=(12, 6))
sns.boxplot(x=billionaires_data['age'])
plt.title('Boxplot of Billionaires Age')
plt.xlabel('Age')
plt.show()

age_outlier_summary

# Select numerical columns for correlation analysis
numerical_data = billionaires_data.select_dtypes(include=['int64', 'float64'])

# Calculate the correlation matrix
correlation_matrix = numerical_data.corr()

# Plotting the heatmap for the correlation matrix
plt.figure(figsize=(12, 10))
sns.heatmap(correlation_matrix, annot=True, cmap='coolwarm', fmt=".2f")
plt.title('Correlation Matrix Heatmap')
plt.show()