利用径向柱图探索星球大战数据

HsuHeinrich

发布于 2025-11-20 11:42:29

260

利用径向柱图探索星球大战数据

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

from matplotlib.lines import Line2D
from matplotlib import font_manager

import warnings
warnings.filterwarnings("ignore", category=RuntimeWarning)

数据探索

以下数据如果有需要的同学可关注公众号HsuHeinrich，回复【数据可视化】自动获取～

# 导入数据
df = pd.read_csv('https://raw.githubusercontent.com/Lisa-Ho/small-data-projects/main/2023/2308-star-wars-scripts/episode1_each_line_of_anakin_clean.csv')

df.head()

image-20240129181218963

《星球大战》阿纳金对话数据： to：对话角色名称 number：数量

基本径向柱图

# 计算每个bar的位置（角度）
x_max = 2*np.pi
df['angular_pos'] = np.linspace(0, x_max, len(df), endpoint=False)

# 自定义颜色
chart_colors = {'bg': '#0C081F', 'QUI-GON': '#F271A7', 'PADME': '#40B8E1', 'OBI-WAN':'#75EAB6',
                'R2D2': '#F4E55E', 'other': '#444A68'}

# 将颜色映射给'to'字段
df['colors'] = df['to'].map(chart_colors)

# 缺失值填充
df['colors'] = df['colors'].fillna(chart_colors['other'])

# 查看调色板
sns.palplot(list(chart_colors.values()))

output_7_0

# 初始化布局
fig, ax = plt.subplots(figsize=(10, 10), subplot_kw={'projection': 'polar'})

# 条形图
ax.bar(df['angular_pos'], df['number'], alpha=1, color=df['colors'] , linewidth=0, width=0.052, zorder=3)

# 自定义轴
max_value = 50
r_offset = -10      
ax.set_rlim(0, max_value)
ax.set_rorigin(r_offset)

# 从顶部开始顺时针绘制条形图
ax.set_theta_zero_location('N')     
ax.set_theta_direction(-1)      

# 标题+著作信息
plt.figtext(0.5,1.03, 'Star Wars Episode I', fontfamily='Mandalore', fontsize=55, ha='center')
plt.figtext(0.5,0.98, 'Each line of Anakin', fontfamily='Ubuntu',fontsize=24, ha='center')
plt.figtext(0.5,0.08, 'Data: jcwieme/data-scripts-star-wars  |  Design: Lisa Hornung',       
            fontfamily='Mandalore',fontsize=8, ha='center', alpha=0.75)

plt.show()

output_8_0

自定义布局

# 自定义缩放y轴
max_value = 50
r_offset = -10
r2 = max_value - r_offset
alpha = r2 - r_offset
v_offset = r_offset**2 / alpha
forward = lambda value: ((value + v_offset) * alpha)**0.5 + r_offset
reverse = lambda radius: (radius - r_offset) ** 2 / alpha - v_offset
ax.set_rlim(0, max_value)
ax.set_rorigin(r_offset)
ax.set_yscale('function', functions=(
    lambda value: np.where(value >= 0, forward(value), value),
    lambda radius: np.where(radius > 0, reverse(radius), radius)))

# 自定义标签
ax.set_rlabel_position(0)
ax.set_yticks([10,20,30,40])
ax.set_yticklabels([10,20,30,40],fontsize=9)

# 自定义网格线
ax.set_thetagrids(angles=[])
ax.grid(visible=True, axis='y', zorder=2, linewidth=0.75)

# 移除边框
ax.spines[:].set_visible(False)

fig

output_10_0

最终的图

# 初始化极坐标图
fig, ax = plt.subplots(figsize=(10, 10), 
                       subplot_kw={'projection': 'polar'})

# 背景色
ax.set_facecolor(chart_colors['bg'])
fig.set_facecolor(chart_colors['bg'])

# 添加条形图
ax.bar(df['angular_pos'], df['number'], alpha=1, color=df['colors'], 
       linewidth=0, width=0.052, zorder=3)

# 从顶部开始顺时针绘制条形图
ax.set_theta_zero_location('N')
ax.set_theta_direction(-1)   

# 自定义y轴缩放
max_value = 50
r_offset = -10
r2 = max_value - r_offset
alpha = r2 - r_offset
v_offset = r_offset**2 / alpha
forward = lambda value: ((value + v_offset) * alpha)**0.5 + r_offset
reverse = lambda radius: (radius - r_offset) ** 2 / alpha - v_offset
ax.set_rlim(0, max_value)
ax.set_rorigin(r_offset)
ax.set_yscale('function', functions=(
    lambda value: np.where(value >= 0, forward(value), value),
    lambda radius: np.where(radius > 0, reverse(radius), radius)))

# 自定义轴标签
ax.set_rlabel_position(0)
ax.set_yticks([10,20,30,40])
ax.set_yticklabels([10,20,30,40],fontsize=9, color='white',alpha=0.35)

# 自定义网格线
ax.set_thetagrids(angles=[])
ax.grid(visible=True, axis='y', zorder=2, color='white',
        linewidth=0.75, alpha=0.2)

# 移除边框
ax.spines[:].set_visible(False)

# 自定义图例
# 添加图例轴
lgd = fig.add_axes([0.75,0.71, 0.15, 0.25]) 

# 图例元素
kw = dict(marker='o', color=chart_colors['bg'], markersize=8, alpha=1, 
          markeredgecolor='None', linewidth=0)
legend_elements =[Line2D([0],[0], 
                          markerfacecolor=chart_colors['PADME'],
                          label='Padme', 
                          **kw), 
                  Line2D([0], [0], 
                         markerfacecolor=chart_colors['QUI-GON'],
                         label='Qui-Gon', 
                         **kw), 
                  Line2D([0], [0], 
                         markerfacecolor=chart_colors['R2D2'], 
                         label='R2D2', 
                         **kw), 
                  Line2D([0], [0], 
                         markerfacecolor=chart_colors['OBI-WAN'], 
                         label='Obi-Wan', 
                         **kw), 
                  Line2D([0], [0], 
                         markerfacecolor=chart_colors['other'], 
                         label='Other', 
                         **kw)] 

# 移除图例边框
L = lgd.legend(frameon=False, handles=legend_elements, loc='center', 
               ncol=1, handletextpad=0.2, labelspacing=1)
plt.setp(L.texts, va='baseline', color='white', size=12, 
         fontfamily='Ubuntu')    
lgd.axis('off')

# 注释
# 新增一个内圆：用于指示绘制方向
circ = fig.add_axes([0.453, 0.435, 0.12, 0.12],polar=True) 
line_angular_pos = df['angular_pos'][1:-5]
line_r = [5] * len(line_angular_pos)

# 绘制开始和结束的标记
circ.plot(line_angular_pos, line_r, zorder=5, color='white', 
          linewidth=0.75, alpha=0.4)
circ.plot(line_angular_pos.to_list()[0], line_r[0], zorder=5, color='white', 
          linewidth=0,marker='o', markersize=3,alpha=0.4)
circ.plot(line_angular_pos.to_list()[-1], line_r[-1], zorder=5, color='white', 
          linewidth=0,marker='>', markersize=3,alpha=0.4)

# 自定义轴：从顶部开始顺时针绘制条形图
circ.set_theta_zero_location('N')
circ.set_theta_direction(-1)  
circ.axis('off')

# 文本注释
ax.annotate('1 line', xy=(0.1, 48), xycoords='data', xytext=(40, 20), 
            textcoords='offset points', 
            fontsize=10, fontfamily='Ubuntu',
            ha='left', va='baseline', 
            annotation_clip=False, 
            color='#ababab',
            arrowprops=dict(arrowstyle='->',edgecolor='#ababab', 
                            connectionstyle='arc3,rad=.5', alpha=0.75))
ax.annotate('Words\nper line', xy=(-0.05, 22), xycoords='data', xytext=(0, 0), 
            textcoords='offset points', 
            fontsize=10, fontfamily='Ubuntu',
            ha='right', va='baseline', 
            annotation_clip=False, 
            color='#ababab')
ax.annotate('', xy=(-0.02, 38), xycoords='data', xytext=(0, -105), 
            textcoords='offset points', 
            fontsize=10, fontfamily='Ubuntu',
            ha='right', va='baseline', 
            annotation_clip=False, 
            color='#ababab',
            arrowprops=dict(arrowstyle='<->',edgecolor='#ababab', linewidth=0.75,
                            connectionstyle='arc3,rad=0', alpha=0.75 ))
lgd.annotate('Talking to', xy=(0.35, 0.78), xycoords='data', xytext=(-18, 14), 
             textcoords='offset points', 
            fontsize=10, fontfamily='Ubuntu',
            ha='right', va='center', 
            annotation_clip=False, 
            color='#ababab',
            arrowprops=dict(arrowstyle='->',edgecolor='#ababab', 
                            connectionstyle='arc3,rad=-.5', alpha=0.75))

# 标题+制作信息
plt.figtext(0.5,1.03, 'Star Wars Episode I', 
            fontfamily='Mandalore', 
            fontsize=55, color='white', ha='center')
plt.figtext(0.5,0.98, 'Each line of Anakin', 
            fontfamily='Ubuntu',
            fontsize=24, color='white', ha='center')
plt.figtext(0.5,0.1, 'Data: jcwieme/data-scripts-star-wars  |  Design: Lisa Hornung', 
            fontfamily='Ubuntu',
            fontsize=8, color='white', ha='center', alpha=0.75)


plt.show()