绘制多索引数据框

Question

我想返回与此类似的情节，但以堆叠的条形图（女性堆叠在男性上方）为例：

第1/2组为年份，男女为来源（例如Facebook，Google Ads等）。

这是我为图表编写的代码：

grouped = filtered.loc[filtered.groupby('anonymous_id')['utm_source']
                                         .apply(lambda x: x.first_valid_index())]\
                            .dropna(subset=['utm_source'])
grouped_counts = grouped.groupby(['year','utm_source'])['Won'].value_counts(normalize=True).mul(100).unstack().plot.bar(stacked=True)

此代码返回以下图形：

以下是我的数据示例：

      year  utm_source  Won                          anonymous_id
0     2018         NaN   No  0000f8ea-3aa6-4423-9247-1d9580d378e1
36    2018  google_ads   No  0008cb14-c833-494f-aa33-2caca8f91ca1
45    2018         NaN   No  000223b0-5278-418d-b9b1-b384024d31a9
46    2018         NaN   No  00037bad-bb90-4836-a9fa-61cbd44e480f
49    2019         NaN  Yes  0000a6a0-00bc-475f-a9e5-9dcbb4309e78
50    2019         NaN  Yes  0000a6a0-00bc-475f-a9e5-9dcbb4309e78
51    2019         NaN  Yes  0000a6a0-00bc-475f-a9e5-9dcbb4309e78
52    2019         NaN  Yes  0000a6a0-00bc-475f-a9e5-9dcbb4309e78
53    2019         NaN  Yes  0000a6a0-00bc-475f-a9e5-9dcbb4309e78
54    2019         NaN  Yes  0000a6a0-00bc-475f-a9e5-9dcbb4309e78