同一图表上的多个 Pandas DataFrame 条形图

Question

如何在同一张图表上绘制多个 DataFrame 条形图？

我想绘制前三名（或前五名）的分数。因为我知道 1st >= 2nd >= 3rd，所以我想将图表中的前三个（或五个）绘制在同一个柱上，而不是将它们分散在三个（或五个）柱上。

视觉效果将与堆叠的条形图完全相同，但条形图并非相互堆叠，而是从底部开始测量。

更新： @DizietAsahi 建议改用堆叠条。我想这是最简单的解决方案。有人可以提供数据框操作代码来获得以下分数的差异吗？

源数据为TID和Score的形式，就像下面的CSV格式的数据一样，已经被过滤掉了，只剩下前3名。对于相同的 TID，原始数据的分数要高得多。这样做的挑战在于，我还需要绘制 MEAN 分数以及前三名。即，我个人认为不可能同时操纵 MEAN 得分和前三名来获得下面的差异。所以任何一种方式都有挑战（对我来说）。

这是一个 CSV 格式的示例数据：

TID,Score
06,510
06,472
06,441
07,630
07,619
07,574
08,617
08,589
08,560
09,610
09,595
09,553
10,593
10,550
10,542
11,442
11,404
11,381

DataFrame 格式（仅适用于多个 DataFrame 条形图的情况。对于使用堆积条形图，为每个 TID 生成一堆随机数作为 Score 就可以了）：

Scores = [
{"TID":07,"ScoreRank":1,"Score":834,"Average":690},
{"TID":07,"ScoreRank":2,"Score":820,"Average":690},
{"TID":07,"ScoreRank":3,"Score":788,"Average":690},
{"TID":08,"ScoreRank":1,"Score":617,"Average":571},
{"TID":08,"ScoreRank":2,"Score":610,"Average":571},
{"TID":08,"ScoreRank":3,"Score":600,"Average":571},
{"TID":09,"ScoreRank":1,"Score":650,"Average":584},
{"TID":09,"ScoreRank":2,"Score":644,"Average":584},
{"TID":09,"ScoreRank":3,"Score":618,"Average":584},
{"TID":10,"ScoreRank":1,"Score":632,"Average":547},
{"TID":10,"ScoreRank":2,"Score":593,"Average":547},
{"TID":10,"ScoreRank":3,"Score":577,"Average":547},
{"TID":11,"ScoreRank":1,"Score":479,"Average":409},
{"TID":11,"ScoreRank":2,"Score":445,"Average":409},
{"TID":11,"ScoreRank":3,"Score":442,"Average":409},
{"TID":12,"ScoreRank":1,"Score":370,"Average":299},
{"TID":12,"ScoreRank":2,"Score":349,"Average":299},
{"TID":12,"ScoreRank":3,"Score":341,"Average":299},
{"TID":13,"ScoreRank":1,"Score":342,"Average":252},
{"TID":13,"ScoreRank":2,"Score":318,"Average":252},
{"TID":13,"ScoreRank":3,"Score":286,"Average":252},
{"TID":14,"ScoreRank":1,"Score":303,"Average":257},
{"TID":14,"ScoreRank":2,"Score":292,"Average":257},
{"TID":14,"ScoreRank":3,"Score":288,"Average":257},
{"TID":15,"ScoreRank":1,"Score":312,"Average":242},
{"TID":15,"ScoreRank":2,"Score":276,"Average":242},
{"TID":15,"ScoreRank":3,"Score":264,"Average":242},
{"TID":16,"ScoreRank":1,"Score":421,"Average":369},
{"TID":16,"ScoreRank":2,"Score":403,"Average":369},
{"TID":16,"ScoreRank":3,"Score":398,"Average":369},
{"TID":17,"ScoreRank":1,"Score":479,"Average":418},
{"TID":17,"ScoreRank":2,"Score":466,"Average":418},
{"TID":17,"ScoreRank":3,"Score":455,"Average":418},
{"TID":18,"ScoreRank":1,"Score":554,"Average":463},
{"TID":18,"ScoreRank":2,"Score":521,"Average":463},
{"TID":18,"ScoreRank":3,"Score":520,"Average":463}]
df = pandas.DataFrame(Scores)

谢谢

Answer 1

这就是你要找的吗？

import pandas
import matplotlib.pyplot as plt
import numpy as np
Scores = [
{"TID":7,"ScoreRank":1,"Score":834,"Average":690},
{"TID":7,"ScoreRank":2,"Score":820,"Average":690},
{"TID":7,"ScoreRank":3,"Score":788,"Average":690},
{"TID":8,"ScoreRank":1,"Score":617,"Average":571},
{"TID":8,"ScoreRank":2,"Score":610,"Average":571},
{"TID":8,"ScoreRank":3,"Score":600,"Average":571},
{"TID":9,"ScoreRank":1,"Score":650,"Average":584},
{"TID":9,"ScoreRank":2,"Score":644,"Average":584},
{"TID":9,"ScoreRank":3,"Score":618,"Average":584},
{"TID":10,"ScoreRank":1,"Score":632,"Average":547},
{"TID":10,"ScoreRank":2,"Score":593,"Average":547},
{"TID":10,"ScoreRank":3,"Score":577,"Average":547},
{"TID":11,"ScoreRank":1,"Score":479,"Average":409},
{"TID":11,"ScoreRank":2,"Score":445,"Average":409},
{"TID":11,"ScoreRank":3,"Score":442,"Average":409},
{"TID":12,"ScoreRank":1,"Score":370,"Average":299},
{"TID":12,"ScoreRank":2,"Score":349,"Average":299},
{"TID":12,"ScoreRank":3,"Score":341,"Average":299},
{"TID":13,"ScoreRank":1,"Score":342,"Average":252},
{"TID":13,"ScoreRank":2,"Score":318,"Average":252},
{"TID":13,"ScoreRank":3,"Score":286,"Average":252},
{"TID":14,"ScoreRank":1,"Score":303,"Average":257},
{"TID":14,"ScoreRank":2,"Score":292,"Average":257},
{"TID":14,"ScoreRank":3,"Score":288,"Average":257},
{"TID":15,"ScoreRank":1,"Score":312,"Average":242},
{"TID":15,"ScoreRank":2,"Score":276,"Average":242},
{"TID":15,"ScoreRank":3,"Score":264,"Average":242},
{"TID":16,"ScoreRank":1,"Score":421,"Average":369},
{"TID":16,"ScoreRank":2,"Score":403,"Average":369},
{"TID":16,"ScoreRank":3,"Score":398,"Average":369},
{"TID":17,"ScoreRank":1,"Score":479,"Average":418},
{"TID":17,"ScoreRank":2,"Score":466,"Average":418},
{"TID":17,"ScoreRank":3,"Score":455,"Average":418},
{"TID":18,"ScoreRank":1,"Score":554,"Average":463},
{"TID":18,"ScoreRank":2,"Score":521,"Average":463},
{"TID":18,"ScoreRank":3,"Score":520,"Average":463}]

df = pandas.DataFrame(Scores)
f, ax1 = plt.subplots(1, figsize=(10,5))
bar_width = 0.75
bar_l = [i+1 for i in range(len(np.unique(df['TID'])))]
tick_pos = [i+(bar_width/2) for i in bar_l]
ax1.bar(bar_l,
        df['Score'][df['ScoreRank'] == 1],
        width=bar_width,
        label='Rank1',
        alpha=0.5,
        color='#eaff0a')

ax1.bar(bar_l,

        df['Score'][df['ScoreRank'] == 2],
        width=bar_width,
        label='Rank2',
        alpha=0.5,
        color='#939393')

ax1.bar(bar_l,
        df['Score'][df['ScoreRank'] == 3],
        width=bar_width,
        label='Rank3',
        alpha=0.5,
        color='#e29024')

ax1.bar(bar_l,
        df['Average'][df['ScoreRank'] == 3],
        width=bar_width,
        label='Average',
        alpha=0.5,
        color='#FF0000')

plt.xticks(tick_pos, np.unique(df['TID']))
ax1.set_ylabel("Score")
ax1.set_xlabel("TID")
plt.legend(loc='upper right')
plt.xlim([min(tick_pos)-bar_width, max(tick_pos)+bar_width])
plt.show()

结果：