Data Visualization Comparison

Comparing matplotlib, seaborn, plotnine, and altair

Introduction

This document compares four popular Python visualization libraries using the Gapminder dataset. Each visualization type is implemented in:

matplotlib: The foundational plotting library
seaborn: Statistical visualization built on matplotlib
plotnine: Grammar of graphics (ggplot2-style)
altair: Declarative visualization based on Vega-Lite

Code

gap.head()

	country	continent	year	lifeExp	pop	gdpPercap	pop_millions
0	Afghanistan	Asia	1952	28.801	8425333	779.445314	8.425333
1	Afghanistan	Asia	1957	30.332	9240934	820.853030	9.240934
2	Afghanistan	Asia	1962	31.997	10267083	853.100710	10.267083
3	Afghanistan	Asia	1967	34.020	11537966	836.197138	11.537966
4	Afghanistan	Asia	1972	36.088	13079460	739.981106	13.079460

Static vs Interactive: Visual Encoding Burden

A key advantage of interactive visualizations is reducing visual encoding burden. In static plots, every dimension you want to communicate must be encoded visually (color, shape, size, alpha, line style). This quickly becomes overwhelming. Interactive plots can offload dimensions to tooltips, keeping the visual clean.

Multi-dimensional Data: 6 Variables at Once

Show country-level data with: continent, income group (high/low GDP), population size, growth rate, country name, and exact values. Static plots need to encode all of this visually.

Code

# Create a complex dataset with multiple dimensions
gap_complex = gap_2007.copy()
gap_complex['income_group'] = pd.cut(
    gap_complex['gdpPercap'],
    bins=[0, 2000, 10000, 50000],
    labels=['Low', 'Middle', 'High']
)
gap_complex['growth'] = gap_complex['country'].map(
    gap[gap['year'].isin([1952, 2007])].groupby('country').apply(
        lambda x: (x[x['year']==2007]['lifeExp'].values[0] - x[x['year']==1952]['lifeExp'].values[0])
        if len(x) == 2 else 0,
        include_groups=False
    )
)
gap_complex['growth_cat'] = pd.cut(
    gap_complex['growth'],
    bins=[-100, 10, 20, 100],
    labels=['Slow (<10yr)', 'Moderate (10-20yr)', 'Fast (>20yr)']
)

Code

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

markers = {'Low': 'v', 'Middle': 's', 'High': '^'}
alphas = {'Slow (<10yr)': 0.3, 'Moderate (10-20yr)': 0.6, 'Fast (>20yr)': 1.0}
colors = {'Africa': '#e41a1c', 'Americas': '#377eb8', 'Asia': '#4daf4a',
          'Europe': '#984ea3', 'Oceania': '#ff7f00'}

for _, row in gap_complex.iterrows():
    ax.scatter(
        row['gdpPercap'], row['lifeExp'],
        c=colors.get(row['continent'], 'gray'),
        marker=markers.get(row['income_group'], 'o'),
        s=row['pop_millions'] * 2,
        alpha=alphas.get(row['growth_cat'], 0.5),
        edgecolors='black',
        linewidths=0.5
    )

# Complex legend construction
from matplotlib.lines import Line2D
legend_elements = []

# Continent colors
for cont, color in colors.items():
    legend_elements.append(Line2D([0], [0], marker='o', color='w',
                                   markerfacecolor=color, markersize=8, label=cont))
legend_elements.append(Line2D([0], [0], color='w', label=''))  # spacer

# Income shapes
for income, marker in markers.items():
    legend_elements.append(Line2D([0], [0], marker=marker, color='w',
                                   markerfacecolor='gray', markersize=8,
                                   label=f'Income: {income}'))
legend_elements.append(Line2D([0], [0], color='w', label=''))  # spacer

# Growth alpha
for growth, alpha in alphas.items():
    legend_elements.append(Line2D([0], [0], marker='o', color='w',
                                   markerfacecolor='gray', markersize=8,
                                   alpha=alpha, label=f'Growth: {growth}'))

ax.legend(handles=legend_elements, loc='lower right', fontsize=8,
          title='Color=Continent, Shape=Income,\nAlpha=Growth, Size=Population')
ax.set_xscale('log')
ax.set_xlabel('GDP per Capita (log scale)')
ax.set_ylabel('Life Expectancy')
ax.set_title('6 Dimensions Encoded Visually\n(Country names not shown - would add more clutter)')
plt.tight_layout()
plt.show()

Code

alt.Chart(gap_complex).mark_circle(size=80, opacity=0.7).encode(
    x=alt.X('gdpPercap:Q', scale=alt.Scale(type='log'), title='GDP per Capita'),
    y=alt.Y('lifeExp:Q', title='Life Expectancy'),
    color=alt.Color('continent:N', title='Continent'),
    tooltip=[
        alt.Tooltip('country:N', title='Country'),
        alt.Tooltip('continent:N', title='Continent'),
        alt.Tooltip('income_group:N', title='Income Group'),
        alt.Tooltip('growth_cat:N', title='Life Exp Growth (1952-2007)'),
        alt.Tooltip('growth:Q', title='Years Gained', format='.1f'),
        alt.Tooltip('gdpPercap:Q', title='GDP/capita', format='$,.0f'),
        alt.Tooltip('lifeExp:Q', title='Life Expectancy', format='.1f'),
        alt.Tooltip('pop_millions:Q', title='Population (M)', format='.1f')
    ]
).properties(
    title='Same 6 Dimensions: Clean Visual, Rich Tooltips',
    width=550,
    height=400
).interactive()

Same six dimensions: color=continent, everything else in tooltip. Hover to explore.

Dense Time Series: 20+ Overlapping Lines

Life expectancy trends for all European countries. In static plots, distinguishing 30 lines requires extensive visual encoding.

Code

fig, ax = plt.subplots(figsize=(12, 7))

europe = gap[gap['continent'] == 'Europe']
countries = europe['country'].unique()

# Try to make lines distinguishable with color, linestyle, marker
linestyles = ['-', '--', '-.', ':']
markers = ['o', 's', '^', 'v', 'D', 'p', 'h', '*']
cmap = plt.cm.get_cmap('tab20')

for i, country in enumerate(countries):
    data = europe[europe['country'] == country]
    ax.plot(data['year'], data['lifeExp'],
            color=cmap(i % 20),
            linestyle=linestyles[i % 4],
            marker=markers[i % 8],
            markersize=4,
            linewidth=1.5,
            alpha=0.7,
            label=country)

# Legend is a mess
ax.legend(bbox_to_anchor=(1.02, 1), loc='upper left', fontsize=6, ncol=2)
ax.set_xlabel('Year')
ax.set_ylabel('Life Expectancy')
ax.set_title('European Life Expectancy Trends\n(Which line is Turkey? Slovenia? Good luck.)')
plt.tight_layout()
plt.show()

/tmp/ipykernel_511434/3704084044.py:9: MatplotlibDeprecationWarning: The get_cmap function was deprecated in Matplotlib 3.7 and will be removed in 3.11. Use ``matplotlib.colormaps[name]`` or ``matplotlib.colormaps.get_cmap()`` or ``pyplot.get_cmap()`` instead.

30 countries need 30 distinguishable visual encodings. Even with varied colors, line styles, and markers, it’s nearly impossible to identify specific countries.

Code

europe = gap[gap['continent'] == 'Europe']

# Selection for highlighting
highlight = alt.selection_point(on='pointerover', fields=['country'], nearest=True)

base = alt.Chart(europe).encode(
    x=alt.X('year:O', title='Year'),
    y=alt.Y('lifeExp:Q', title='Life Expectancy', scale=alt.Scale(domain=[55, 85])),
    color=alt.condition(
        highlight,
        alt.Color('country:N', legend=None),
        alt.value('lightgray')
    ),
    opacity=alt.condition(highlight, alt.value(1), alt.value(0.3)),
    tooltip=[
        alt.Tooltip('country:N', title='Country'),
        alt.Tooltip('year:O', title='Year'),
        alt.Tooltip('lifeExp:Q', title='Life Expectancy', format='.1f'),
        alt.Tooltip('pop_millions:Q', title='Population (M)', format='.1f'),
        alt.Tooltip('gdpPercap:Q', title='GDP/capita', format='$,.0f')
    ]
)

lines = base.mark_line(strokeWidth=2).add_params(highlight)
points = base.mark_circle(size=50).add_params(highlight)

(lines + points).properties(
    title='Hover to Highlight Any Country',
    width=550,
    height=400
)

Same 30 countries, but hover reveals identity. No legend clutter needed.

Scatterplot

Relationship between GDP per capita and life expectancy in 2007.

Code

fig, ax = plt.subplots(figsize=(10, 6))

for continent, data in gap_2007.groupby('continent'):
    ax.scatter(data['gdpPercap'], data['lifeExp'],
               s=data['pop_millions'], alpha=0.6, label=continent)

ax.set_xlabel('GDP per Capita')
ax.set_ylabel('Life Expectancy')
ax.set_title('GDP vs Life Expectancy (2007)')
ax.legend(title='Continent')
ax.set_xscale('log')
plt.tight_layout()
plt.show()

Code

fig, ax = plt.subplots(figsize=(10, 6))
sns.scatterplot(data=gap_2007, x='gdpPercap', y='lifeExp',
                hue='continent', size='pop_millions',
                sizes=(20, 500), alpha=0.6, ax=ax)
ax.set_xscale('log')
ax.set_xlabel('GDP per Capita')
ax.set_ylabel('Life Expectancy')
ax.set_title('GDP vs Life Expectancy (2007)')
plt.tight_layout()
plt.show()

Code

(
    p9.ggplot(gap_2007, p9.aes(x='gdpPercap', y='lifeExp',
                                color='continent', size='pop_millions'))
    + p9.geom_point(alpha=0.6)
    + p9.scale_x_log10()
    + p9.labs(title='GDP vs Life Expectancy (2007)',
              x='GDP per Capita', y='Life Expectancy')
    + p9.theme_minimal()
)

Code

alt.Chart(gap_2007).mark_circle().encode(
    x=alt.X('gdpPercap:Q', scale=alt.Scale(type='log'), title='GDP per Capita'),
    y=alt.Y('lifeExp:Q', title='Life Expectancy'),
    color='continent:N',
    size=alt.Size('pop_millions:Q', title='Population (M)'),
    tooltip=['country', 'gdpPercap', 'lifeExp', 'pop_millions']
).properties(
    title='GDP vs Life Expectancy (2007)',
    width=500,
    height=350
).interactive()

Line Plot

Life expectancy over time for selected countries.

Code

fig, ax = plt.subplots(figsize=(10, 6))

for country, data in gap_5countries.groupby('country'):
    ax.plot(data['year'], data['lifeExp'], marker='o', label=country)

ax.set_xlabel('Year')
ax.set_ylabel('Life Expectancy')
ax.set_title('Life Expectancy Over Time')
ax.legend(title='Country')
plt.tight_layout()
plt.show()

Code

fig, ax = plt.subplots(figsize=(10, 6))
sns.lineplot(data=gap_5countries, x='year', y='lifeExp',
             hue='country', marker='o', ax=ax)
ax.set_xlabel('Year')
ax.set_ylabel('Life Expectancy')
ax.set_title('Life Expectancy Over Time')
plt.tight_layout()
plt.show()

Code

(
    p9.ggplot(gap_5countries, p9.aes(x='year', y='lifeExp', color='country'))
    + p9.geom_line()
    + p9.geom_point()
    + p9.labs(title='Life Expectancy Over Time',
              x='Year', y='Life Expectancy')
    + p9.theme_minimal()
)

Code

alt.Chart(gap_5countries).mark_line(point=True).encode(
    x=alt.X('year:O', title='Year'),
    y=alt.Y('lifeExp:Q', title='Life Expectancy'),
    color='country:N',
    tooltip=['country', 'year', 'lifeExp']
).properties(
    title='Life Expectancy Over Time',
    width=500,
    height=350
).interactive()

Bar Plot

Average life expectancy by continent in 2007.

Code

continent_avg = gap_2007.groupby('continent')['lifeExp'].mean().reset_index()
continent_avg.columns = ['continent', 'avg_lifeExp']

Code

fig, ax = plt.subplots(figsize=(8, 5))
ax.bar(continent_avg['continent'], continent_avg['avg_lifeExp'],
       color='steelblue', edgecolor='black')
ax.set_xlabel('Continent')
ax.set_ylabel('Average Life Expectancy')
ax.set_title('Average Life Expectancy by Continent (2007)')
plt.tight_layout()
plt.show()

Code

fig, ax = plt.subplots(figsize=(8, 5))
sns.barplot(data=continent_avg, x='continent', y='avg_lifeExp',
            color='steelblue', edgecolor='black', ax=ax)
ax.set_xlabel('Continent')
ax.set_ylabel('Average Life Expectancy')
ax.set_title('Average Life Expectancy by Continent (2007)')
plt.tight_layout()
plt.show()

Code

(
    p9.ggplot(continent_avg, p9.aes(x='continent', y='avg_lifeExp'))
    + p9.geom_col(fill='steelblue', color='black')
    + p9.labs(title='Average Life Expectancy by Continent (2007)',
              x='Continent', y='Average Life Expectancy')
    + p9.theme_minimal()
)

Code

alt.Chart(continent_avg).mark_bar(color='steelblue').encode(
    x=alt.X('continent:N', title='Continent'),
    y=alt.Y('avg_lifeExp:Q', title='Average Life Expectancy'),
    tooltip=['continent', 'avg_lifeExp']
).properties(
    title='Average Life Expectancy by Continent (2007)',
    width=400,
    height=300
)

Histogram

Distribution of life expectancy in 2007.

Code

fig, ax = plt.subplots(figsize=(8, 5))
ax.hist(gap_2007['lifeExp'], bins=20, color='steelblue', edgecolor='black')
ax.set_xlabel('Life Expectancy')
ax.set_ylabel('Frequency')
ax.set_title('Distribution of Life Expectancy (2007)')
plt.tight_layout()
plt.show()

Code

fig, ax = plt.subplots(figsize=(8, 5))
sns.histplot(data=gap_2007, x='lifeExp', bins=20,
             color='steelblue', edgecolor='black', ax=ax)
ax.set_xlabel('Life Expectancy')
ax.set_ylabel('Frequency')
ax.set_title('Distribution of Life Expectancy (2007)')
plt.tight_layout()
plt.show()

Code

(
    p9.ggplot(gap_2007, p9.aes(x='lifeExp'))
    + p9.geom_histogram(bins=20, fill='steelblue', color='black')
    + p9.labs(title='Distribution of Life Expectancy (2007)',
              x='Life Expectancy', y='Frequency')
    + p9.theme_minimal()
)

Code

alt.Chart(gap_2007).mark_bar(color='steelblue').encode(
    x=alt.X('lifeExp:Q', bin=alt.Bin(maxbins=20), title='Life Expectancy'),
    y=alt.Y('count()', title='Frequency'),
    tooltip=['count()']
).properties(
    title='Distribution of Life Expectancy (2007)',
    width=400,
    height=300
)

Cleveland Dot Plot

Life expectancy by country in Europe (2007), ordered by value.

Code

europe_2007 = gap[(gap['continent'] == 'Europe') & (gap['year'] == 2007)].copy()
europe_2007 = europe_2007.sort_values('lifeExp')

Code

fig, ax = plt.subplots(figsize=(8, 10))
ax.scatter(europe_2007['lifeExp'], europe_2007['country'],
           color='steelblue', s=50)
ax.hlines(y=europe_2007['country'], xmin=europe_2007['lifeExp'].min() - 1,
          xmax=europe_2007['lifeExp'], color='gray', linewidth=0.5, alpha=0.5)
ax.set_xlabel('Life Expectancy')
ax.set_ylabel('Country')
ax.set_title('Life Expectancy in Europe (2007)')
plt.tight_layout()
plt.show()

Code

fig, ax = plt.subplots(figsize=(8, 10))
sns.stripplot(data=europe_2007, x='lifeExp', y='country',
              color='steelblue', size=8, ax=ax)
ax.set_xlabel('Life Expectancy')
ax.set_ylabel('Country')
ax.set_title('Life Expectancy in Europe (2007)')
plt.tight_layout()
plt.show()

Code

(
    p9.ggplot(europe_2007, p9.aes(x='lifeExp',
                                   y='pd.Categorical(country, categories=europe_2007["country"])'))
    + p9.geom_point(color='steelblue', size=3)
    + p9.geom_segment(p9.aes(x=europe_2007['lifeExp'].min() - 1,
                              xend='lifeExp',
                              yend='pd.Categorical(country, categories=europe_2007["country"])'),
                       color='gray', alpha=0.5)
    + p9.labs(title='Life Expectancy in Europe (2007)',
              x='Life Expectancy', y='Country')
    + p9.theme_minimal()
)

Code

alt.Chart(europe_2007).mark_circle(color='steelblue', size=80).encode(
    x=alt.X('lifeExp:Q', title='Life Expectancy'),
    y=alt.Y('country:N', sort=alt.EncodingSortField(field='lifeExp', order='ascending'),
            title='Country'),
    tooltip=['country', 'lifeExp']
).properties(
    title='Life Expectancy in Europe (2007)',
    width=400,
    height=500
)

Grouped Bar Plot

Average life expectancy by continent across selected years.

Code

gap_years = gap[gap['year'].isin([1957, 1977, 1997, 2007])]
continent_year_avg = gap_years.groupby(['continent', 'year'])['lifeExp'].mean().reset_index()
continent_year_avg.columns = ['continent', 'year', 'avg_lifeExp']
continent_year_avg['year'] = continent_year_avg['year'].astype(str)

Code

fig, ax = plt.subplots(figsize=(12, 6))

continents = continent_year_avg['continent'].unique()
years = continent_year_avg['year'].unique()
x = np.arange(len(continents))
width = 0.2

for i, year in enumerate(years):
    data = continent_year_avg[continent_year_avg['year'] == year]
    ax.bar(x + i * width, data['avg_lifeExp'], width, label=year)

ax.set_xlabel('Continent')
ax.set_ylabel('Average Life Expectancy')
ax.set_title('Average Life Expectancy by Continent and Year')
ax.set_xticks(x + width * 1.5)
ax.set_xticklabels(continents)
ax.legend(title='Year')
plt.tight_layout()
plt.show()

Code

fig, ax = plt.subplots(figsize=(12, 6))
sns.barplot(data=continent_year_avg, x='continent', y='avg_lifeExp',
            hue='year', ax=ax)
ax.set_xlabel('Continent')
ax.set_ylabel('Average Life Expectancy')
ax.set_title('Average Life Expectancy by Continent and Year')
plt.tight_layout()
plt.show()

Code

(
    p9.ggplot(continent_year_avg, p9.aes(x='continent', y='avg_lifeExp', fill='year'))
    + p9.geom_col(position='dodge')
    + p9.labs(title='Average Life Expectancy by Continent and Year',
              x='Continent', y='Average Life Expectancy')
    + p9.theme_minimal()
)

Code

alt.Chart(continent_year_avg).mark_bar().encode(
    x=alt.X('year:N', title=None),
    y=alt.Y('avg_lifeExp:Q', title='Average Life Expectancy'),
    color=alt.Color('year:N', title='Year'),
    column=alt.Column('continent:N', title='Continent'),
    tooltip=['continent', 'year', 'avg_lifeExp']
).properties(
    width=100,
    title='Average Life Expectancy by Continent and Year'
)

Choropleth Map

Life expectancy by country in 2007.

Code

import geopandas as gpd

# Load world boundaries from Natural Earth
world = gpd.read_file('https://naciscdn.org/naturalearth/110m/cultural/ne_110m_admin_0_countries.zip')

# Prepare country name mapping for join (gapminder -> Natural Earth NAME)
country_mapping = {
    'United States': 'United States of America',
    'Congo, Dem. Rep.': 'Dem. Rep. Congo',
    'Congo, Rep.': 'Congo',
    'Korea, Rep.': 'South Korea',
    'Korea, Dem. Rep.': 'North Korea',
    'Yemen, Rep.': 'Yemen',
    'Czech Republic': 'Czechia',
    'Slovak Republic': 'Slovakia',
    "Cote d'Ivoire": "Ivory Coast",
    'West Bank and Gaza': 'Palestine',
    'Bosnia and Herzegovina': 'Bosnia and Herz.',
    'Central African Republic': 'Central African Rep.',
    'Dominican Republic': 'Dominican Rep.',
    'Equatorial Guinea': 'Eq. Guinea',
    'South Sudan': 'S. Sudan',
    'Trinidad and Tobago': 'Trinidad and Tobago'
}

gap_2007_map = gap_2007.copy()
gap_2007_map['NAME'] = gap_2007_map['country'].replace(country_mapping)

# Merge with world geometries
world_data = world.merge(gap_2007_map, on='NAME', how='left')

Code

fig, ax = plt.subplots(figsize=(15, 8))
world_data.plot(column='lifeExp', ax=ax, legend=True,
                legend_kwds={'label': 'Life Expectancy', 'shrink': 0.6},
                missing_kwds={'color': 'lightgrey'},
                cmap='YlGnBu')
ax.set_title('Life Expectancy by Country (2007)')
ax.axis('off')
plt.tight_layout()
plt.show()

Code

# Seaborn does not have native choropleth support
# Using matplotlib with seaborn styling
sns.set_style("whitegrid")
fig, ax = plt.subplots(figsize=(15, 8))
world_data.plot(column='lifeExp', ax=ax, legend=True,
                legend_kwds={'label': 'Life Expectancy', 'shrink': 0.6},
                missing_kwds={'color': 'lightgrey'},
                cmap='YlGnBu')
ax.set_title('Life Expectancy by Country (2007)')
ax.axis('off')
plt.tight_layout()
plt.show()
sns.reset_defaults()

Code

# plotnine has limited choropleth support via geom_map
# Using a workaround with geom_polygon
(
    p9.ggplot()
    + p9.geom_map(data=world_data, mapping=p9.aes(fill='lifeExp'))
    + p9.scale_fill_cmap('YlGnBu', na_value='lightgrey', name='Life Expectancy')
    + p9.labs(title='Life Expectancy by Country (2007)')
    + p9.theme_void()
    + p9.theme(figure_size=(15, 8))
)

Code

from vega_datasets import data as vega_data

# ISO 3166-1 numeric codes mapping to gapminder country names
iso_numeric_to_country = {
    4: 'Afghanistan', 8: 'Albania', 12: 'Algeria', 24: 'Angola', 32: 'Argentina',
    36: 'Australia', 40: 'Austria', 48: 'Bahrain', 50: 'Bangladesh', 56: 'Belgium',
    204: 'Benin', 68: 'Bolivia', 70: 'Bosnia and Herzegovina', 72: 'Botswana',
    76: 'Brazil', 100: 'Bulgaria', 854: 'Burkina Faso', 108: 'Burundi',
    116: 'Cambodia', 120: 'Cameroon', 124: 'Canada', 140: 'Central African Republic',
    148: 'Chad', 152: 'Chile', 156: 'China', 170: 'Colombia', 174: 'Comoros',
    180: 'Congo, Dem. Rep.', 178: 'Congo, Rep.', 188: 'Costa Rica',
    384: "Cote d'Ivoire", 191: 'Croatia', 192: 'Cuba', 203: 'Czech Republic',
    208: 'Denmark', 262: 'Djibouti', 214: 'Dominican Republic', 218: 'Ecuador',
    818: 'Egypt', 222: 'El Salvador', 226: 'Equatorial Guinea', 232: 'Eritrea',
    231: 'Ethiopia', 246: 'Finland', 250: 'France', 266: 'Gabon', 270: 'Gambia',
    276: 'Germany', 288: 'Ghana', 300: 'Greece', 320: 'Guatemala', 324: 'Guinea',
    624: 'Guinea-Bissau', 332: 'Haiti', 340: 'Honduras', 344: 'Hong Kong, China',
    348: 'Hungary', 352: 'Iceland', 356: 'India', 360: 'Indonesia', 364: 'Iran',
    368: 'Iraq', 372: 'Ireland', 376: 'Israel', 380: 'Italy', 388: 'Jamaica',
    392: 'Japan', 400: 'Jordan', 404: 'Kenya', 408: 'Korea, Dem. Rep.',
    410: 'Korea, Rep.', 414: 'Kuwait', 422: 'Lebanon', 426: 'Lesotho',
    430: 'Liberia', 434: 'Libya', 450: 'Madagascar', 454: 'Malawi', 458: 'Malaysia',
    466: 'Mali', 478: 'Mauritania', 480: 'Mauritius', 484: 'Mexico', 496: 'Mongolia',
    499: 'Montenegro', 504: 'Morocco', 508: 'Mozambique', 104: 'Myanmar',
    516: 'Namibia', 524: 'Nepal', 528: 'Netherlands', 554: 'New Zealand',
    558: 'Nicaragua', 562: 'Niger', 566: 'Nigeria', 578: 'Norway', 512: 'Oman',
    586: 'Pakistan', 591: 'Panama', 600: 'Paraguay', 604: 'Peru', 608: 'Philippines',
    616: 'Poland', 620: 'Portugal', 630: 'Puerto Rico', 642: 'Romania',
    646: 'Rwanda', 678: 'Sao Tome and Principe', 682: 'Saudi Arabia', 686: 'Senegal',
    688: 'Serbia', 694: 'Sierra Leone', 702: 'Singapore', 703: 'Slovak Republic',
    705: 'Slovenia', 706: 'Somalia', 710: 'South Africa', 724: 'Spain',
    144: 'Sri Lanka', 729: 'Sudan', 748: 'Swaziland', 752: 'Sweden',
    756: 'Switzerland', 760: 'Syria', 158: 'Taiwan', 834: 'Tanzania',
    764: 'Thailand', 768: 'Togo', 780: 'Trinidad and Tobago', 788: 'Tunisia',
    792: 'Turkey', 800: 'Uganda', 826: 'United Kingdom', 840: 'United States',
    858: 'Uruguay', 862: 'Venezuela', 704: 'Vietnam', 275: 'West Bank and Gaza',
    887: 'Yemen, Rep.', 894: 'Zambia', 716: 'Zimbabwe'
}

# Create lookup dataframe with id column
gap_2007_iso = gap_2007.copy()
country_to_iso = {v: k for k, v in iso_numeric_to_country.items()}
gap_2007_iso['id'] = gap_2007_iso['country'].map(country_to_iso)

countries = alt.topo_feature(vega_data.world_110m.url, 'countries')

# Layer approach: grey background for all countries, colored overlay for data
background = alt.Chart(countries).mark_geoshape(
    fill='lightgrey',
    stroke='white',
    strokeWidth=0.5
).project('naturalEarth1')

choropleth = alt.Chart(countries).mark_geoshape(
    stroke='white',
    strokeWidth=0.5
).transform_lookup(
    lookup='id',
    from_=alt.LookupData(gap_2007_iso, 'id', ['lifeExp', 'country'])
).encode(
    color=alt.Color('lifeExp:Q', scale=alt.Scale(scheme='yellowgreenblue'),
                    title='Life Expectancy', legend=alt.Legend(orient='bottom')),
    tooltip=[
        alt.Tooltip('country:N', title='Country'),
        alt.Tooltip('lifeExp:Q', title='Life Expectancy', format='.1f')
    ]
).project('naturalEarth1')

(background + choropleth).properties(
    title='Life Expectancy by Country (2007)',
    width=700,
    height=400
)

Summary

Feature	matplotlib	seaborn	plotnine	altair
Syntax	Imperative	Imperative	Declarative (ggplot)	Declarative (Vega-Lite)
Learning curve	Moderate	Low	Low (if familiar with R)	Low
Customization	Very high	High	High	Moderate
Interactivity	Limited	Limited	None	Built-in
Statistical plots	Manual	Built-in	Built-in	Manual
Choropleths	Via geopandas	Via geopandas	Limited	Built-in
Output	Static	Static	Static	Interactive

Each library has its strengths:

matplotlib: Maximum control and customization
seaborn: Best for quick statistical visualizations
plotnine: Familiar grammar of graphics for R users
altair: Best for interactive web visualizations

Appendix: Project Configuration

pyproject.toml

[project]
name = "viz-comparison"
version = "0.1.0"
description = "Add your description here"
readme = "README.md"
requires-python = ">=3.11"
dependencies = [
    "altair>=6.0.0",
    "gapminder>=0.1",
    "geopandas>=1.1.2",
    "jupyter>=1.1.1",
    "matplotlib>=3.10.8",
    "nbclient>=0.10.4",
    "nbformat>=5.10.4",
    "plotnine>=0.15.3",
    "seaborn>=0.13.2",
    "setuptools>=82.0.0",
    "vega-datasets>=0.9.0",
]