Pandasデータ可視化: Matplotlib統合をマスターし、グローバルな洞察を得る

            pip install pandas matplotlib
            

              
                Copy
              
            
          

            import pandas as pd
import matplotlib.pyplot as plt
import numpy as np # Often useful for generating sample data
            

              
                Copy
              
            
          

            data = {
    'Quarter': ['Q1', 'Q2', 'Q3', 'Q4', 'Q1', 'Q2', 'Q3', 'Q4'],
    'Year': [2022, 2022, 2022, 2022, 2023, 2023, 2023, 2023],
    'North America Sales (USD)': [150, 160, 175, 180, 190, 200, 210, 220],
    'Europe Sales (USD)': [120, 130, 140, 135, 145, 155, 165, 170],
    'Asia Sales (USD)': [100, 115, 130, 150, 160, 175, 190, 200],
    'Africa Sales (USD)': [50, 55, 60, 65, 70, 75, 80, 85],
    'Latin America Sales (USD)': [80, 85, 90, 95, 100, 105, 110, 115]
}
df = pd.DataFrame(data)
df['Date'] = pd.to_datetime(df['Year'].astype(str) + df['Quarter'].str.replace('Q', '-Q'))
df = df.set_index('Date')
print(df.head())
            

              
                Copy
              
            
          

            df[['North America Sales (USD)', 'Europe Sales (USD)', 'Asia Sales (USD)']].plot(
    kind='line',
    figsize=(12, 6),
    title='Regional Sales Performance Over Time (2022-2023)',
    xlabel='Date',
    ylabel='Sales (USD Millions)',
    grid=True
)
plt.show()
            

              
                Copy
              
            
          

            yearly_sales = df.groupby('Year')[['North America Sales (USD)', 'Europe Sales (USD)', 'Asia Sales (USD)', 'Africa Sales (USD)', 'Latin America Sales (USD)']].sum()
yearly_sales.plot(
    kind='bar',
    figsize=(14, 7),
    title='Total Yearly Sales by Region (2022 vs 2023)',
    ylabel='Total Sales (USD Millions)',
    rot=45, # Rotate x-axis labels for better readability
    width=0.8
)
plt.tight_layout() # Adjust layout to prevent labels from overlapping
plt.show()
            

              
                Copy
              
            
          

            # Let's create some dummy data for "Customer Satisfaction Scores" (out of 100) from two global regions
np.random.seed(42)
customer_satisfaction_na = np.random.normal(loc=85, scale=10, size=500)
customer_satisfaction_eu = np.random.normal(loc=78, scale=12, size=500)

satisfaction_df = pd.DataFrame({
    'North America': customer_satisfaction_na,
    'Europe': customer_satisfaction_eu
})

satisfaction_df.plot(
    kind='hist',
    bins=20, # Number of bins
    alpha=0.7, # Transparency
    figsize=(10, 6),
    title='Distribution of Customer Satisfaction Scores by Region',
    xlabel='Satisfaction Score',
    ylabel='Frequency',
    grid=True,
    legend=True
)
plt.show()
            

              
                Copy
              
            
          

            # Let's imagine we have data on 'Marketing Spend' and 'Sales' for various product launches globally
scatter_data = {
    'Marketing Spend (USD)': np.random.uniform(50, 500, 100),
    'Sales (USD)': np.random.uniform(100, 1000, 100),
    'Region': np.random.choice(['NA', 'EU', 'Asia', 'Africa', 'LA'], 100)
}
scatter_df = pd.DataFrame(scatter_data)
# Introduce some correlation
scatter_df['Sales (USD)'] = scatter_df['Sales (USD)'] + scatter_df['Marketing Spend (USD)'] * 1.5

scatter_df.plot(
    kind='scatter',
    x='Marketing Spend (USD)',
    y='Sales (USD)',
    figsize=(10, 6),
    title='Global Marketing Spend vs. Sales Performance',
    s=scatter_df['Marketing Spend (USD)'] / 5, # Marker size proportional to spend
    c='blue', # Color of markers
    alpha=0.6,
    grid=True
)
plt.show()
            

              
                Copy
              
            
          

            # Let's use our satisfaction_df for box plots
satisfaction_df.plot(
    kind='box',
    figsize=(8, 6),
    title='Customer Satisfaction Score Distribution by Region',
    ylabel='Satisfaction Score',
    grid=True
)
plt.show()
            

              
                Copy
              
            
          

            # Let's consider monthly energy consumption (in KWh) for a company's global operations
energy_data = {
    'Month': pd.to_datetime(pd.date_range(start='2023-01', periods=12, freq='M')),
    'North America (KWh)': np.random.randint(1000, 1500, 12) + np.arange(12)*20,
    'Europe (KWh)': np.random.randint(800, 1200, 12) + np.arange(12)*15,
    'Asia (KWh)': np.random.randint(1200, 1800, 12) + np.arange(12)*25,
}
energy_df = pd.DataFrame(energy_data).set_index('Month')

energy_df.plot(
    kind='area',
    stacked=True, # Stack the areas
    figsize=(12, 6),
    title='Monthly Global Energy Consumption by Region (KWh)',
    xlabel='Month',
    ylabel='Total Energy Consumption (KWh)',
    alpha=0.8,
    grid=True
)
plt.show()
            

              
                Copy
              
            
          

            # Generate the Pandas plot and capture the Axes object
ax = df[['North America Sales (USD)', 'Europe Sales (USD)', 'Asia Sales (USD)']].plot(
    kind='line',
    figsize=(12, 7),
    title='Regional Sales Performance Over Time (2022-2023)',
    xlabel='Date',
    ylabel='Sales (USD Millions)',
    grid=True
)

# Now, use Matplotlib's Axes methods for further customization
ax.set_facecolor('#f0f0f0') # Light grey background for the plotting area
ax.spines['top'].set_visible(False) # Remove top spine
ax.spines['right'].set_visible(False) # Remove right spine
ax.tick_params(axis='x', rotation=30) # Rotate x-tick labels
ax.tick_params(axis='y', labelcolor='darkgreen') # Change y-tick label color

# Add a specific annotation for a significant point
# Let's say we had a major marketing campaign start in Q3 2023 in Asia
asia_q3_2023_sales = df.loc['2023-09-30', 'Asia Sales (USD)'] # Assuming Q3 ends Sep 30
ax.annotate(f'Asia Campaign: {asia_q3_2023_sales:.0f}M USD',
            xy=('2023-09-30', asia_q3_2023_sales),
            xytext=('2023-05-01', asia_q3_2023_sales + 30), # Offset text from point
            arrowprops=dict(facecolor='black', shrink=0.05),
            fontsize=10,
            color='darkred',
            bbox=dict(boxstyle="round,pad=0.3", fc="yellow", ec="darkgrey", lw=0.5, alpha=0.9)) # Double quotes here need escaping

# Improve legend placement
ax.legend(title='Region', bbox_to_anchor=(1.05, 1), loc='upper left')

# Adjust layout to make room for the legend
plt.tight_layout(rect=[0, 0, 0.85, 1])

# Save the figure with high resolution, suitable for global reports
plt.savefig('regional_sales_performance_enhanced.png', dpi=300, bbox_inches='tight')
plt.show()
            

              
                Copy
              
            
          

            # Let's visualize the distribution of sales for North America and Europe separately
fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(14, 6))

# Plot North America sales distribution on the first Axes
df['North America Sales (USD)'].plot(
    kind='hist',
    ax=axes[0],
    bins=10,
    alpha=0.7,
    color='skyblue',
    edgecolor='black'
)
axes[0].set_title('North America Sales Distribution')
axes[0].set_xlabel('Sales (USD Millions)')
axes[0].set_ylabel('Frequency')
axes[0].grid(axis='y', linestyle='--', alpha=0.7)

# Plot Europe sales distribution on the second Axes
df['Europe Sales (USD)'].plot(
    kind='hist',
    ax=axes[1],
    bins=10,
    alpha=0.7,
    color='lightcoral',
    edgecolor='black'
)
axes[1].set_title('Europe Sales Distribution')
axes[1].set_xlabel('Sales (USD Millions)')
axes[1].set_ylabel('') # Remove redundant Y-label as it's shared
axes[1].grid(axis='y', linestyle='--', alpha=0.7)

fig.suptitle('Sales Distribution Comparison (2022-2023)', fontsize=16) # Overall figure title
plt.tight_layout(rect=[0, 0.03, 1, 0.95]) # Adjust layout for suptitle
plt.show()
            

              
                Copy
              
            
          

            # Sample data: Quarterly GDP growth rates (percentage) for different continents
gdp_data = {
    'Quarter': pd.to_datetime(pd.date_range(start='2021-01', periods=12, freq='Q')),
    'North America GDP Growth (%)': np.random.uniform(0.5, 2.0, 12),
    'Europe GDP Growth (%)': np.random.uniform(0.2, 1.8, 12),
    'Asia GDP Growth (%)': np.random.uniform(1.0, 3.5, 12),
    'Africa GDP Growth (%)': np.random.uniform(0.0, 2.5, 12),
    'Latin America GDP Growth (%)': np.random.uniform(-0.5, 2.0, 12)
}
gdp_df = pd.DataFrame(gdp_data).set_index('Quarter')

fig, ax = plt.subplots(figsize=(15, 8))

# Pandas plot for the initial line chart
gdp_df.plot(
    kind='line',
    ax=ax,
    marker='o', # Add markers for data points
    linewidth=2,
    alpha=0.8
)

# Matplotlib customizations
ax.set_title('Quarterly GDP Growth Rates by Continent (2021-2023)', fontsize=16, fontweight='bold')
ax.set_xlabel('Quarter', fontsize=12)
ax.set_ylabel('GDP Growth (%)', fontsize=12)
ax.grid(True, linestyle='--', alpha=0.6)
ax.axhline(y=0, color='red', linestyle=':', linewidth=1.5, label='Zero Growth Line') # Add a zero line
ax.legend(title='Continent', loc='upper left', bbox_to_anchor=(1, 1))

# Highlight a specific period (e.g., a global economic downturn period)
ax.axvspan(pd.to_datetime('2022-04-01'), pd.to_datetime('2022-09-30'), color='gray', alpha=0.2, label='Economic Slowdown Period')

# Customizing Y-axis tick labels to add percentage sign
from matplotlib.ticker import PercentFormatter
ax.yaxis.set_major_formatter(PercentFormatter())

plt.tight_layout(rect=[0, 0, 0.88, 1]) # Adjust layout for legend
plt.show()
            

              
                Copy
              
            
          

            # Sample data: Population distribution by age group for two countries
population_data = {
    'Age Group': ['0-14', '15-29', '30-44', '45-59', '60-74', '75+'],
    'Country A (Millions)': [20, 25, 30, 22, 15, 8],
    'Country B (Millions)': [15, 20, 25, 28, 20, 12]
}
pop_df = pd.DataFrame(population_data).set_index('Age Group')

fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(16, 7), sharey=True) # Share Y-axis for easier comparison

# Plot for Country A
pop_df[['Country A (Millions)']].plot(
    kind='barh', # Horizontal bar chart
    ax=axes[0],
    color='skyblue',
    edgecolor='black',
    legend=False
)
axes[0].set_title('Country A Population Distribution', fontsize=14)
axes[0].set_xlabel('Population (Millions)', fontsize=12)
axes[0].set_ylabel('Age Group', fontsize=12)
axes[0].grid(axis='x', linestyle='--', alpha=0.7)
axes[0].invert_xaxis() # Make bars extend left

# Plot for Country B
pop_df[['Country B (Millions)']].plot(
    kind='barh',
    ax=axes[1],
    color='lightcoral',
    edgecolor='black',
    legend=False
)
axes[1].set_title('Country B Population Distribution', fontsize=14)
axes[1].set_xlabel('Population (Millions)', fontsize=12)
axes[1].set_ylabel('') # Remove redundant Y-label as it's shared
axes[1].grid(axis='x', linestyle='--', alpha=0.7)

fig.suptitle('Comparative Population Age Distribution (Global Example)', fontsize=16, fontweight='bold')
plt.tight_layout(rect=[0, 0, 1, 0.95])
plt.show()
            

              
                Copy
              
            
          

            # Sample data: Hypothetical CO2 emissions and GDP per capita for various countries
# Data for 20 global sample countries (simplified)
countries = ['USA', 'CHN', 'IND', 'GBR', 'DEU', 'FRA', 'JPN', 'BRA', 'CAN', 'AUS',
             'MEX', 'IDN', 'NGA', 'EGY', 'ZAF', 'ARG', 'KOR', 'ITA', 'ESP', 'RUS']
np.random.seed(42)
co2_emissions = np.random.uniform(2, 20, len(countries)) * 10 # in metric tons per capita
gdp_per_capita = np.random.uniform(5000, 70000, len(countries))
# Introduce a positive correlation
co2_emissions = co2_emissions + (gdp_per_capita / 5000) * 0.5
co2_emissions = np.clip(co2_emissions, 5, 25) # Ensure reasonable range

env_df = pd.DataFrame({
    'Country': countries,
    'CO2 Emissions (metric tons per capita)': co2_emissions,
    'GDP per Capita (USD)': gdp_per_capita
})

fig, ax = plt.subplots(figsize=(12, 8))

# Pandas scatter plot
env_df.plot(
    kind='scatter',
    x='GDP per Capita (USD)',
    y='CO2 Emissions (metric tons per capita)',
    ax=ax,
    s=env_df['GDP per Capita (USD)'] / 500, # Marker size based on GDP (as a proxy for economic scale)
    alpha=0.7,
    edgecolor='black',
    color='darkgreen'
)

# Matplotlib customizations
ax.set_title('CO2 Emissions vs. GDP per Capita for Global Economies', fontsize=16, fontweight='bold')
ax.set_xlabel('GDP per Capita (USD)', fontsize=12)
ax.set_ylabel('CO2 Emissions (metric tons per capita)', fontsize=12)
ax.grid(True, linestyle=':', alpha=0.5)

# Add country labels for specific points
for i, country in enumerate(env_df['Country']):
    if country in ['USA', 'CHN', 'IND', 'DEU', 'NGA']: # Label a few interesting countries
        ax.text(env_df['GDP per Capita (USD)'].iloc[i] + 500, # Offset x
                env_df['CO2 Emissions (metric tons per capita)'].iloc[i] + 0.5, # Offset y
                country,
                fontsize=9,
                color='darkblue',
                fontweight='bold')

plt.tight_layout()
plt.show()
            

              
                Copy