Python绘制高级环形柱状图(含代码)

引言

在科研数据可视化中，传统柱状图往往只能呈现“高低对比”，但当数据包含多组分类、多子类别结构时，线性排列的柱状图很容易显得冗长、信息密集却缺乏层次感。环形柱状图（Circular Barplot）通过极坐标布局，将“分组—子类—数值”三层信息整合在一个紧凑的结构中，不仅提升视觉张力，也更利于整体结构的理解。

本文基于 Python 极坐标绘图，构建两种不同复杂度的环形柱状图：第一种强调结构完整性，第二种强调表达强化与视觉优化。两者在分组数量、柱宽分配以及数值呈现方式上各有侧重，适用于不同类型的科研表达场景。

1、12组 × 4子类 —— 强调结构完整性

第一幅图采用 12 个主分组（A–L），每组包含 4 个子类别（如 a1–a4、b1–b4 等），这一图的设计重点在于保持所有分组结构的完整呈现，子类别标签沿极坐标自动旋转，保证可读性，由于分组数量较多，柱宽相对较窄，整体更偏向“结构地图”的功能。

2、6组 × 6子类 —— 强调数值表达与视觉强化

第二幅图将分组数量减少至 6 个（A–F），同时每组扩展为 6 个子类别。在减少主分组数量后，每组所占极角显著增加，并标注具体数值，比上一幅图更直观判断不同类别的走势，适用于需要展示具体数值差异的结果图。

具体Python如下：

import numpy as np

import pandas as pd

import matplotlib.pyplot as plt

from pathlib import Path

import string

# =========================

# 0) 参数区

# =========================

OUT = Path("ring_bar_out5")

OUT.mkdir(exist_ok=True)

DPI = 600

FONT = "Times New Roman"

# Fig1：12组，每组4柱

N_GROUPS_FIG1 = 12

N_SUB_FIG1 = 4

# Fig2：6组，每组6柱

N_GROUPS_FIG2 = 6

N_SUB_FIG2 = 6

# 环形几何（0~1 归一化半径）

INNER_R = 0.22

BAR_MAX = 0.80

ARC_R = 1.08

ARC_LW = 2.8

GROUP_LABEL_R = 1.15

# 组间留白（空槽）

GAP_SLOTS_FIG1 = 2

GAP_SLOTS_FIG2 = 3

# Fig2：标数值

ANNOTATE_VALUES_FIG2 = True

# Fig1 配色（12 组）

PALETTE_FIG1 = [

    "#1f77b4", "#2ca02c", "#ff7f0e", "#d62728", "#9467bd", "#e377c2",

    "#8c564b", "#17becf", "#bcbd22", "#7f7f7f", "#4c72b0", "#dd8452"

]

# Fig2 配色（6 组，换一套更稳重的）

PALETTE_FIG2 = [

    "#264653", "#2A9D8F", "#E9C46A", "#F4A261", "#E76F51", "#6D597A"

]

# ======== 关键：让 Fig2 的柱子更“长”（值更大） ========

# 1) 全局抬升：所有组整体更大

FIG2_GLOBAL_BOOST = 0.18   # 想更长就 0.25；想保守就 0.10

# 2) 指定组额外抬升：例如 A / F 以前太短，这里再加一点

FIG2_GROUP_EXTRA_BOOST = {"A": 0.18, "F": 0.18}  # 想更长就 0.25

# Fig2 扇区间隔（度）：越小每组越宽、每柱越宽

FIG2_GAP_DEG = 3.0  # 2.0 更宽；4.0 更留白

# =========================

# 1) 工具函数

# =========================

def make_groups(n):

    return list(string.ascii_uppercase[:n])

def build_layout(groups, n_per_group, gap_slots):

    n_groups = len(groups)

    total_slots = n_groups * n_per_group + (n_groups - 1) * gap_slots

    slot_angles = np.linspace(0, 2*np.pi, total_slots, endpoint=False)

    bar_angles = []

    group_spans = []  # (g, start_slot, end_slot)

    slot = 0

    for i, g in enumerate(groups):

        start = slot

        for _ in range(n_per_group):

            bar_angles.append(slot_angles[slot])

            slot += 1

        end = slot - 1

        group_spans.append((g, start, end))

        if i != n_groups - 1:

            slot += gap_slots

    return np.array(bar_angles), slot_angles, group_spans, total_slots

def rot_align(angle_rad):

    a = angle_rad % (2*np.pi)

    rot = np.degrees(a)

    ha = "left"

    if np.pi/2 < a < 3*np.pi/2:

        rot += 180

        ha = "right"

    return rot, ha

def sublabels_for_group(g, n_sub):

    prefix = g.lower()

    return [f"{prefix}{i}" for i in range(1, n_sub + 1)]

def gen_demo(groups, n_sub, seed=7):

    rng = np.random.default_rng(seed)

    rows = []

    for g in groups:

        base = 0.18 + 0.70 * rng.random()

        offsets = np.linspace(-0.08, 0.16, n_sub)

        offsets = offsets + rng.normal(0, 0.03, size=n_sub)

        vals = np.clip(base + offsets + rng.normal(0, 0.07, size=n_sub), 0.03, 1.0)

        for lab, v in zip(sublabels_for_group(g, n_sub), vals):

            rows.append([g, lab, float(v)])

    return pd.DataFrame(rows, columns=["group", "sub", "value"])

def plot_circular(df, groups, n_sub, palette, gap_slots, out_png, title, annotate_values=False):

    plt.rcParams.update({

        "font.family": FONT,

        "font.size": 11,

        "figure.facecolor": "white",

        "axes.facecolor": "white",

    })

    df = df.copy()

    df["group"] = pd.Categorical(df["group"], categories=groups, ordered=True)

    df["sub_num"] = df["sub"].str.extract(r"(\d+)$").astype(int)

    df = df.sort_values(["group", "sub_num"]).reset_index(drop=True)

    bar_angles, slot_angles, group_spans, total_slots = build_layout(groups, n_sub, gap_slots)

    vals = df["value"].to_numpy()

    r_top = INNER_R + vals * (BAR_MAX - INNER_R)

    group_color = {g: palette[i % len(palette)] for i, g in enumerate(groups)}

    bar_colors = df["group"].map(lambda g: group_color[str(g)]).to_numpy()

    subs = df["sub"].astype(str).tolist()

    fig = plt.figure(figsize=(8.8, 8.8))

    ax = plt.subplot(111, polar=True)

    ax.set_theta_direction(-1)

    ax.set_theta_offset(np.pi/2)

    ax.set_xticks([])

    ax.set_yticks([0.2, 0.4, 0.6, 0.8, 1.0])

    ax.set_yticklabels(["0.2", "0.4", "0.6", "0.8", "1.0"])

    ax.grid(alpha=0.30, linestyle="--", linewidth=0.6)

    ax.spines["polar"].set_visible(False)

    # 组底色扇区（淡）

    for g, start, end in group_spans:

        a1 = slot_angles[start]

        a2 = slot_angles[end] + (2*np.pi / total_slots) * 0.86

        if a2 < a1:

            a2 += 2*np.pi

        mid = (a1 + a2) / 2

        width = (a2 - a1)

        ax.bar(mid, 1.02 - INNER_R, width=width, bottom=INNER_R,

               color=group_color[g], alpha=0.08, linewidth=0)

    # 柱子

    width = (2*np.pi / total_slots) * 0.90

    ax.bar(bar_angles, r_top - INNER_R, width=width, bottom=INNER_R,

           color=bar_colors, edgecolor="white", linewidth=1.2, alpha=0.97)

    # 子柱标签

    for ang, rr, s in zip(bar_angles, r_top, subs):

        rot, ha = rot_align(ang)

        ax.text(ang, rr + 0.055, s,

                rotation=rot, rotation_mode="anchor",

                ha=ha, va="center",

                fontsize=9.8, color="0.35")

    # 数值标注

    if annotate_values:

        for ang, rr, v in zip(bar_angles, r_top, vals):

            ax.text(ang, rr - 0.05, f"{v:.2f}",

                    ha="center", va="center",

                    fontsize=8.5, color="black", alpha=0.95)

    # 外圈弧线 + 组名

    for g, start, end in group_spans:

        a1 = slot_angles[start]

        a2 = slot_angles[end] + (2*np.pi / total_slots) * 0.86

        if a2 < a1:

            a2 += 2*np.pi

        tt = np.linspace(a1, a2, 180)

        ax.plot(tt, np.full_like(tt, ARC_R), color=group_color[g], linewidth=ARC_LW)

        mid = (a1 + a2) / 2

        rot, ha = rot_align(mid)

        ax.text(mid % (2*np.pi), GROUP_LABEL_R, g,

                rotation=rot, rotation_mode="anchor",

                ha=ha, va="center",

                fontsize=13, fontweight="bold",

                color=group_color[g],

                bbox=dict(boxstyle="round,pad=0.22", fc="white", ec="none", alpha=0.92))

    ax.set_title(title, pad=24, fontsize=15, fontweight="bold")

    fig.savefig(out_png, dpi=DPI, bbox_inches="tight")

    plt.close(fig)

def plot_fig2_wider(df, groups, n_sub, palette, out_png, title):

    plt.rcParams.update({

        "font.family": FONT,

        "font.size": 11,

        "figure.facecolor": "white",

        "axes.facecolor": "white",

    })

    df = df.copy()

    df["group"] = pd.Categorical(df["group"], categories=groups, ordered=True)

    df["sub_num"] = df["sub"].str.extract(r"(\d+)$").astype(int)

    df = df.sort_values(["group", "sub_num"]).reset_index(drop=True)

    gap_deg = np.deg2rad(FIG2_GAP_DEG)

    group_span = (2*np.pi - gap_deg * len(groups)) / len(groups)

    group_color = {g: palette[i % len(palette)] for i, g in enumerate(groups)}

    fig = plt.figure(figsize=(8.8, 8.8))

    ax = plt.subplot(111, polar=True)

    ax.set_theta_direction(-1)

    ax.set_theta_offset(np.pi/2)

    ax.set_xticks([])

    ax.set_yticks([0.2, 0.4, 0.6, 0.8, 1.0])

    ax.set_yticklabels(["0.2", "0.4", "0.6", "0.8", "1.0"])

    ax.grid(alpha=0.30, linestyle="--", linewidth=0.6)

    ax.spines["polar"].set_visible(False)

    theta0 = 0.0

    for g in groups:

        sub = df[df["group"] == g].copy()

        vals = sub["value"].to_numpy()

        r_top = INNER_R + vals * (BAR_MAX - INNER_R)

        a1 = theta0

        a2 = theta0 + group_span

        mid = (a1 + a2) / 2

        # 组底色扇区（淡）

        ax.bar(mid, 1.02 - INNER_R, width=group_span, bottom=INNER_R,

               color=group_color[g], alpha=0.08, linewidth=0)

        centers = np.linspace(a1 + group_span/(2*n_sub), a2 - group_span/(2*n_sub), n_sub)

        bar_w = 0.92 * (group_span / n_sub)

        ax.bar(centers, r_top - INNER_R, width=bar_w, bottom=INNER_R,

               color=group_color[g], edgecolor="white", linewidth=1.2, alpha=0.97)

        subs = sub["sub"].astype(str).tolist()

        for ang, rr, lab in zip(centers, r_top, subs):

            rot, ha = rot_align(ang)

            ax.text(ang, rr + 0.055, lab,

                    rotation=rot, rotation_mode="anchor",

                    ha=ha, va="center",

                    fontsize=10, color="0.35")

        if ANNOTATE_VALUES_FIG2:

            for ang, rr, v in zip(centers, r_top, vals):

                ax.text(ang, rr - 0.05, f"{v:.2f}",

                        ha="center", va="center",

                        fontsize=9.2, color="black", alpha=0.95)

        tt = np.linspace(a1, a2, 220)

        ax.plot(tt, np.full_like(tt, ARC_R), color=group_color[g], linewidth=ARC_LW)

        rot, ha = rot_align(mid)

        ax.text(mid % (2*np.pi), GROUP_LABEL_R, g,

                rotation=rot, rotation_mode="anchor",

                ha=ha, va="center",

                fontsize=13, fontweight="bold",

                color=group_color[g],

                bbox=dict(boxstyle="round,pad=0.22", fc="white", ec="none", alpha=0.92))

        theta0 = a2 + gap_deg

    ax.set_title(title, pad=24, fontsize=15, fontweight="bold")

    fig.savefig(out_png, dpi=DPI, bbox_inches="tight")

    plt.close(fig)

# =========================

# 2) 生成两张图

# =========================

# Fig1：12组 A~L，每组4柱，不标数值

groups1 = make_groups(N_GROUPS_FIG1)

df1 = gen_demo(groups1, N_SUB_FIG1, seed=11)

plot_circular(

    df1, groups1, N_SUB_FIG1, PALETTE_FIG1, GAP_SLOTS_FIG1,

    OUT / "Fig1_12groups_4sub.png",

    title="Fig1  Circular barplot (12 groups; a1~l4 labels)",

    annotate_values=False

)

# Fig2：6组 A~F，每组6柱，柱更宽 + 值更大（柱更长）+ 标数值

groups2 = make_groups(N_GROUPS_FIG2)

df2 = gen_demo(groups2, N_SUB_FIG2, seed=29)

# ======== 关键：只对 Fig2 的数据抬高，让柱子更长 ========

df2["value"] = np.clip(df2["value"] + FIG2_GLOBAL_BOOST, 0.0, 1.0)

for gg, boost in FIG2_GROUP_EXTRA_BOOST.items():

    m = df2["group"] == gg

    df2.loc[m, "value"] = np.clip(df2.loc[m, "value"] + boost, 0.0, 1.0)

plot_fig2_wider(

    df2, groups2, N_SUB_FIG2, PALETTE_FIG2,

    OUT / "Fig2_6groups_6sub_values_WIDER_BOOST.png",

    title="Fig2  Circular barplot (6 groups; wider sectors + boosted values)"

)

print("[ok] saved to:", OUT.resolve())

print(" -", OUT / "Fig1_12groups_4sub.png")

print(" -", OUT / "Fig2_6groups_6sub_values_WIDER_BOOST.png")

特别声明：

以上代码与文案均为网上资料整合而成，仅供广大同行们参考学习，如有侵权请联系删除。

如有其他需要，欢迎关注我的咸鱼号:pfc小姐姐