Python Pandas：绘制100％堆积图问题

Question

我得到了一个数据帧df5，下面是我从read_csv中读到的表，

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％堆积条形图，每个销售总产品和总利润，其中x轴是周日，标签是不同的类别。

我销售的总产品代码是

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')

我得到的图表看起来没什么我需要的。它不是100堆叠而且图例是已售出的总产品，而不是类别中的不同类别。

有人可以帮忙吗？感谢。

此致 Lobbie

Answer 1

最简单的方法是使用您关注的值制作数据透视表。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：

您可以单独为Total_Profit执行相同的操作。