python - 绘制 100% 堆积图问题

Question

我从 read_csv 中读取了下表，得到了一个数据帧 df5，

Week_Days,Category,Total_Products_Sold,Total_Profit
0.Monday,A,3221,9999.53
0.Monday,B,1038,26070.33
0.Monday,C,699,13779.56
0.Monday,E,3055,18157.26
0.Monday,F,47569,215868.15
0.Monday,G,2348,23695.25
0.Monday,H,6,57
0.Monday,I,14033,64594.24
0.Monday,J,13876,47890.91
0.Monday,K,3878,14119.74
0.Monday,L,243,2649.6
0.Monday,M,2992,16757.38
1.Tuesday,A,2839,8864.78
1.Tuesday,B,1013,26254.69
1.Tuesday,C,656,13206.98
1.Tuesday,E,2696,15872.45
1.Tuesday,F,43039,197621.18
1.Tuesday,G,2107,21048.72
1.Tuesday,H,3,17
1.Tuesday,I,12297,56942.99
1.Tuesday,J,12095,40724.2
1.Tuesday,K,3418,12551.26
1.Tuesday,L,243,2520.3
1.Tuesday,M,2375,13268.28
2.Wednesday,A,2936,9119.93
2.Wednesday,B,1061,26927.86
2.Wednesday,C,634,10424.05
2.Wednesday,E,2835,16627.35
2.Wednesday,F,46128,218014.59
2.Wednesday,G,1986,19173.64
4.Friday,H,24,233
4.Friday,I,17576,81648.75
4.Friday,J,16468,55820.9
4.Friday,K,4294,16603.39
4.Friday,L,440,4258.51
4.Friday,M,3600,20142.44
5.Saturday,A,4658,15051.13
5.Saturday,B,1492,38236.07
5.Saturday,C,1057,15449.7
5.Saturday,E,5335,29904.96
5.Saturday,F,79925,362120.61
5.Saturday,G,4324,44088.79
5.Saturday,H,26,933
5.Saturday,I,22688,106313.86
5.Saturday,J,21882,74725.11
5.Saturday,K,5402,20875.84
5.Saturday,L,458,4692.84
5.Saturday,M,4896,27769.68
6.Sunday,A,3429,11310.1
6.Sunday,B,1104,27282.99
6.Sunday,C,1051,11567.08
6.Sunday,E,3913,22740.63
6.Sunday,F,56048,259105.03
6.Sunday,G,3224,32528.39
6.Sunday,H,21,749
6.Sunday,I,15853,74876.77
6.Sunday,J,16072,55259.76
6.Sunday,K,4383,16058.36
6.Sunday,L,327,3348.82
6.Sunday,M,3551,20814.05

我想绘制 2 个 100% 堆叠条形图，分别表示总销售产品和总利润，其中 x 轴是工作日，标签是不同的类别。

我的 Total Products Sold 代码是

df5 = df5.set_index(['Week_Days', 'Category'])
df5 = df5.div(df5.sum(1), axis=0)
ax = df5[['Total_Products_Sold']].plot(kind='bar', stacked=True, width = 0.3, figsize=(20, 10), colormap="RdBu")
patches, labels = ax.get_legend_handles_labels()
ax.legend(bbox_to_anchor=(1.1, 1.0))
ax.set_xlabel('Week Days')
ax.set_ylabel('Products Sold')

我返回的图表看起来并不需要。它不是 100 个堆叠的，图例是 Total Products Sold 而不是 Category 中的不同类别。

有人可以帮忙吗？谢谢。

问候，大堂

Answer 1

最简单的方法是用你关心的值制作一个数据透视表。尝试这样的事情:

tps = df5.pivot_table(values=['Total_Products_Sold'], 
                      index='Week_Days',
                      columns='Category',
                      aggfunc='sum')

tps = tps.div(tps.sum(1), axis=0)
tps.plot(kind='bar', stacked=True)

对我来说，这会产生以下结果:

您可以分别对 Total_Profit 执行相同的操作。