A bar chart arranged in a circle with bars radiating outward from the center. Each bar's angle represents a category (often direction) and length represents magnitude. Commonly used as a wind rose for meteorological data.
""" anyplot.ai
polar-bar: Polar Bar Chart (Wind Rose)
Library: plotnine 0.15.4 | Python 3.13.13
Quality: 90/100 | Updated: 2026-05-13
"""
import os
import numpy as np
import pandas as pd
from plotnine import (
aes,
coord_fixed,
element_blank,
element_rect,
element_text,
geom_path,
geom_polygon,
geom_segment,
geom_text,
ggplot,
labs,
scale_fill_manual,
scale_x_continuous,
scale_y_continuous,
theme,
)
# Theme tokens
THEME = os.getenv("ANYPLOT_THEME", "light")
PAGE_BG = "#FAF8F1" if THEME == "light" else "#1A1A17"
INK = "#1A1A17" if THEME == "light" else "#F0EFE8"
INK_SOFT = "#4A4A44" if THEME == "light" else "#B8B7B0"
# Okabe-Ito palette
IMPRINT = ["#009E73", "#C475FD", "#4467A3", "#BD8233", "#AE3030", "#2ABCCD", "#954477"]
# Data - Wind direction frequencies (8 compass directions)
directions = ["N", "NE", "E", "SE", "S", "SW", "W", "NW"]
direction_angles = [0, 45, 90, 135, 180, 225, 270, 315] # Degrees from North, clockwise
frequencies = [15, 8, 12, 5, 18, 22, 10, 7]
# Create polygons for each bar (wedge shape)
bar_half_width = 18 # degrees, slightly less than 22.5 for visual gap
bar_rows = []
bar_id = 0
for direction, angle, freq in zip(directions, direction_angles, frequencies, strict=True):
start_angle = angle - bar_half_width
end_angle = angle + bar_half_width
# Center point
points = [(0, 0)]
# Arc points at the outer radius
arc_angles = np.linspace(start_angle, end_angle, 10)
for a in arc_angles:
# Convert from compass (N=0, CW) to math (E=0, CCW)
theta = np.radians(90 - a)
x = freq * np.cos(theta)
y = freq * np.sin(theta)
points.append((x, y))
# Close back to center
points.append((0, 0))
# Add all points to dataframe
for i, (x, y) in enumerate(points):
bar_rows.append({"x": x, "y": y, "direction": direction, "bar_id": bar_id, "order": i})
bar_id += 1
bar_df = pd.DataFrame(bar_rows)
# Create circular gridlines (concentric circles at magnitude intervals)
grid_rows = []
grid_angles = np.linspace(0, 2 * np.pi, 101)
max_radius = max(frequencies) + 5
grid_radii = [5, 10, 15, 20, 25]
for radius in grid_radii:
if radius <= max_radius:
for angle in grid_angles:
grid_rows.append({"x": radius * np.cos(angle), "y": radius * np.sin(angle), "radius": radius})
grid_df = pd.DataFrame(grid_rows)
# Create radial spokes (8 compass directions)
spoke_rows = []
for deg in direction_angles:
angle = np.radians(90 - deg)
spoke_rows.append(
{"x1": 0, "y1": 0, "x2": (max_radius + 2) * np.cos(angle), "y2": (max_radius + 2) * np.sin(angle)}
)
spoke_df = pd.DataFrame(spoke_rows)
# Create compass direction labels
label_rows = []
label_radius = max_radius + 5
for deg, lbl in zip(direction_angles, directions, strict=True):
angle = np.radians(90 - deg)
label_rows.append({"label": lbl, "x": label_radius * np.cos(angle), "y": label_radius * np.sin(angle)})
label_df = pd.DataFrame(label_rows)
# Create radius labels (frequency values) - positioned along NNE axis
radius_labels = []
label_angle = np.radians(90 - 22.5) # NNE direction
for r in [5, 10, 15, 20]:
if r <= max_radius:
radius_labels.append({"label": f"{r}%", "x": r * np.cos(label_angle) + 1.5, "y": r * np.sin(label_angle)})
radius_label_df = pd.DataFrame(radius_labels)
# Color palette - first direction uses brand green, rest alternate through Okabe-Ito
colors = {
"N": IMPRINT[0],
"NE": IMPRINT[1],
"E": IMPRINT[2],
"SE": IMPRINT[3],
"S": IMPRINT[4],
"SW": IMPRINT[5],
"W": IMPRINT[6],
"NW": IMPRINT[0],
}
# Plot
plot = (
ggplot()
# Circular gridlines (frequency circles)
+ geom_path(
aes(x="x", y="y", group="radius"), data=grid_df, color=INK_SOFT, size=0.5, alpha=0.15, linetype="dashed"
)
# Radial spokes (direction lines)
+ geom_segment(aes(x="x1", y="y1", xend="x2", yend="y2"), data=spoke_df, color=INK_SOFT, size=0.5, alpha=0.15)
# Bar wedges (wind rose bars)
+ geom_polygon(
aes(x="x", y="y", group="bar_id", fill="direction"),
data=bar_df,
color=INK_SOFT,
size=0.5,
alpha=0.85,
show_legend=False,
)
# Compass direction labels
+ geom_text(aes(x="x", y="y", label="label"), data=label_df, size=16, color=INK, fontweight="bold")
# Frequency labels
+ geom_text(aes(x="x", y="y", label="label"), data=radius_label_df, size=10, color=INK_SOFT, ha="left")
# Custom colors for directions
+ scale_fill_manual(values=colors)
# Equal coordinate system for proper circles
+ coord_fixed(ratio=1)
# Axis scaling with padding
+ scale_x_continuous(limits=(-35, 35))
+ scale_y_continuous(limits=(-35, 35))
# Title
+ labs(title="Wind Direction Frequency (%) · polar-bar · plotnine · anyplot.ai")
# Theme with theme-adaptive colors
+ theme(
figure_size=(12, 12),
plot_title=element_text(size=24, ha="center", color=INK),
axis_title=element_blank(),
axis_text=element_blank(),
axis_ticks=element_blank(),
axis_line=element_blank(),
panel_grid_major=element_blank(),
panel_grid_minor=element_blank(),
panel_background=element_rect(fill=PAGE_BG, color=PAGE_BG),
plot_background=element_rect(fill=PAGE_BG, color=PAGE_BG),
)
)
# Save
plot.save(f"plot-{THEME}.png", dpi=300, verbose=False)
Part of Polar Bar Chart (Wind Rose) on anyplot.ai.