A bubble chart extending scatter plots by adding a third dimension through bubble size. Each point's position shows two variables (x, y) while the bubble size represents a third quantitative variable. This visualization is excellent for understanding relationships between three numerical variables simultaneously, revealing patterns that would be hidden in traditional 2D scatter plots.
""" anyplot.ai
bubble-basic: Basic Bubble Chart
Library: bokeh 3.9.0 | Python 3.13.13
Quality: 92/100 | Created: 2026-05-28
"""
import base64
import io
import os
import time
from pathlib import Path
import numpy as np
from bokeh.io import output_file, save
from bokeh.models import BoxAnnotation, ColumnDataSource, HoverTool, Label, LinearColorMapper, Range1d
from bokeh.plotting import figure
from bokeh.transform import transform
from PIL import Image
from selenium import webdriver
from selenium.webdriver.chrome.options import Options
# Theme tokens
THEME = os.getenv("ANYPLOT_THEME", "light")
PAGE_BG = "#FAF8F1" if THEME == "light" else "#1A1A17"
ELEVATED_BG = "#FFFDF6" if THEME == "light" else "#242420"
INK = "#1A1A17" if THEME == "light" else "#F0EFE8"
INK_SOFT = "#4A4A44" if THEME == "light" else "#B8B7B0"
# imprint_seq colormap (single-polarity: brand green → blue)
_t = np.linspace(0, 1, 256)
_c0 = np.array([0x00, 0x9E, 0x73])
_c1 = np.array([0x44, 0x67, 0xA3])
ANYPLOT_SEQ256 = ["#{:02X}{:02X}{:02X}".format(*(_c0 + (_c1 - _c0) * t).round().astype(int)) for t in _t]
# Data — City metrics: population density vs median income, bubble = green space per capita
# Green space inversely correlated with density (denser cities have less green space)
np.random.seed(42)
n_cities = 40
population_density = np.random.uniform(500, 12000, n_cities) # people per km²
median_income = 30 + population_density / 400 + np.random.normal(0, 4, n_cities) # thousands USD
green_space = 60 - (population_density / 12000) * 50 + np.random.normal(0, 4, n_cities)
green_space = np.clip(green_space, 5, 60) # m² per capita
# Area-proportional bubble sizes: size² ∝ data value, so size ∝ sqrt(data)
size_min, size_max = 22, 80
green_norm = (green_space - green_space.min()) / (green_space.max() - green_space.min())
bubble_size = np.sqrt(size_min**2 + (size_max**2 - size_min**2) * green_norm)
color_mapper = LinearColorMapper(palette=ANYPLOT_SEQ256, low=green_space.min(), high=green_space.max())
source = ColumnDataSource(
data={
"density": population_density,
"income": median_income,
"size": bubble_size,
"green_space": green_space,
"density_display": np.round(population_density).astype(int),
"income_display": np.round(median_income, 1),
"green_display": np.round(green_space, 1),
}
)
# Symmetric axis ranges — equal padding both sides; extra top space for legend
x_pad = (population_density.max() - population_density.min()) * 0.07
y_pad = (median_income.max() - median_income.min()) * 0.07
x_start = population_density.min() - x_pad * 1.5
x_end = population_density.max() + x_pad * 1.5
y_start = median_income.min() - y_pad
y_end = median_income.max() + y_pad * 7
x_range = x_end - x_start
y_range = y_end - y_start
# Plot
title = "bubble-basic · python · bokeh · anyplot.ai"
p = figure(
width=3200,
height=1800,
title=title,
x_axis_label="Population Density (people/km²)",
y_axis_label="Median Income (thousands USD)",
toolbar_location=None,
min_border_bottom=160,
min_border_left=180,
min_border_top=110,
min_border_right=50,
)
p.x_range = Range1d(start=x_start, end=x_end)
p.y_range = Range1d(start=y_start, end=y_end)
p.scatter(
x="density",
y="income",
size="size",
source=source,
fill_color=transform("green_space", color_mapper),
fill_alpha=0.65,
line_color=PAGE_BG,
line_width=2,
)
# Hover tool
hover = HoverTool(
tooltips=[
("Density", "@density_display{,} people/km²"),
("Income", "$@income_display{0.0}k"),
("Green Space", "@green_display m²/capita"),
],
mode="mouse",
)
p.add_tools(hover)
# Dashed trend line — guides viewer to the positive density-income correlation
trend_coeffs = np.polyfit(population_density, median_income, 1)
x_trend = np.linspace(x_start, x_end, 100)
y_trend = np.polyval(trend_coeffs, x_trend)
p.line(x=x_trend, y=y_trend, line_color=INK_SOFT, line_dash="dashed", line_width=5, line_alpha=0.5)
# Theme-adaptive chrome
p.background_fill_color = PAGE_BG
p.border_fill_color = PAGE_BG
p.outline_line_color = None
p.outline_line_alpha = 0
p.title.text_font_size = "50pt"
p.title.text_color = INK
p.xaxis.axis_label_text_font_size = "42pt"
p.yaxis.axis_label_text_font_size = "42pt"
p.xaxis.axis_label_text_color = INK
p.yaxis.axis_label_text_color = INK
p.xaxis.major_label_text_font_size = "34pt"
p.yaxis.major_label_text_font_size = "34pt"
p.xaxis.major_label_text_color = INK_SOFT
p.yaxis.major_label_text_color = INK_SOFT
p.xaxis.axis_line_color = INK_SOFT
p.yaxis.axis_line_color = INK_SOFT
p.xaxis.major_tick_line_color = INK_SOFT
p.yaxis.major_tick_line_color = INK_SOFT
p.xaxis.minor_tick_line_color = None
p.yaxis.minor_tick_line_color = None
p.xgrid.grid_line_color = INK
p.ygrid.grid_line_color = INK
p.xgrid.grid_line_alpha = 0.12
p.ygrid.grid_line_alpha = 0.12
# Size legend — anchored above the main data cluster (top region is empty due to correlation)
legend_cx = x_start + x_range * 0.22
legend_top = y_end - y_range * 0.04
y_step = y_range * 0.07
ref_green = [green_space.min(), (green_space.min() + green_space.max()) / 2, green_space.max()]
ref_norm = [(v - green_space.min()) / (green_space.max() - green_space.min()) for v in ref_green]
ref_sizes = [np.sqrt(size_min**2 + (size_max**2 - size_min**2) * n) for n in ref_norm]
ref_labels = [f"{v:.0f} m²/capita" for v in ref_green]
legend_box = BoxAnnotation(
left=legend_cx - x_range * 0.13,
right=legend_cx + x_range * 0.13,
top=legend_top + y_range * 0.01,
bottom=legend_top - y_step * 3.8,
fill_color=ELEVATED_BG,
fill_alpha=0.9,
line_color=INK_SOFT,
line_alpha=0.4,
)
p.add_layout(legend_box)
p.add_layout(
Label(
x=legend_cx,
y=legend_top - y_range * 0.01,
text="Green Space",
text_font_size="30pt",
text_font_style="bold",
text_color=INK,
text_align="center",
)
)
for i, (sz, lbl, gv) in enumerate(zip(ref_sizes, ref_labels, ref_green, strict=True)):
ly = legend_top - y_step * (i + 0.85)
ref_src = ColumnDataSource(data={"x": [legend_cx - x_range * 0.04], "y": [ly], "size": [sz], "green_space": [gv]})
p.scatter(
x="x",
y="y",
size="size",
source=ref_src,
fill_color=transform("green_space", color_mapper),
fill_alpha=0.65,
line_color=PAGE_BG,
line_width=2,
)
p.add_layout(
Label(
x=legend_cx + x_range * 0.01,
y=ly,
text=lbl,
text_font_size="26pt",
text_baseline="middle",
text_color=INK_SOFT,
)
)
# Save HTML
output_file(f"plot-{THEME}.html")
save(p)
# Save PNG via headless Chrome — use captureBeyondViewport so browser chrome
# overhead (~139px) doesn't truncate the canvas height
W, H = 3200, 1800
opts = Options()
for arg in (
"--headless=new",
"--no-sandbox",
"--disable-dev-shm-usage",
"--disable-gpu",
f"--window-size={W},{H}",
"--hide-scrollbars",
):
opts.add_argument(arg)
driver = webdriver.Chrome(options=opts)
driver.set_window_size(W, H)
driver.get(f"file://{Path(f'plot-{THEME}.html').resolve()}")
time.sleep(3)
screenshot = driver.execute_cdp_cmd("Page.captureScreenshot", {"format": "png", "captureBeyondViewport": True})
driver.quit()
Image.open(io.BytesIO(base64.b64decode(screenshot["data"]))).save(f"plot-{THEME}.png")
Part of Basic Bubble Chart on anyplot.ai.