嵌套 Pandas 数据框的 FaceGrid

Question

我需要使用 FaceGrid 对以下数据框进行箱线图绘制，以使网格的每一行都对应于 k_week 给出的特定周，并且列是数据框的列 (Sensitivity, Specificity 等）。每个箱线图应包含 6 个模型（“Sig”、“MRM”、“Mean”等）

下面是单行的例子（我成功搞定了）：

现在我需要使用 FaceGrid 在一个图形上获得这样的 6 行。

数据框：

df.groupby('k_week').head(6)
   Sensitivity  Specificity  Accuracy       PPV       AUC  k_week    model
0     0.609948     0.867516  0.717291  0.776004  0.813680       4      Sig
1     0.586824     0.851647  0.685525  0.757557  0.784805       4      MRM
2     0.581152     0.842510  0.670357  0.749651  0.737269       4     Mean
3     0.473386     0.601827  0.395841  0.556193  0.542679       4    Rmssd
4     0.511344     0.717240  0.499148  0.644086  0.626719       4  MissVal
5     0.454625     0.523924  0.356061  0.508448  0.501661       4     Rand
0     0.792619     0.764423  0.755839  0.777934  0.835871       6      Sig
1     0.722077     0.758113  0.733090  0.740845  0.803949       6      MRM
2     0.633899     0.735877  0.688298  0.687011  0.736115       6     Mean
3     0.437296     0.613582  0.510095  0.529108  0.544014       6    Rmssd
4     0.517309     0.740986  0.647588  0.633803  0.653312       6  MissVal
5     0.439582     0.532452  0.477898  0.502504  0.499943       6     Rand
0     0.716681     0.778103  0.762680  0.747469  0.826501       8      Sig
1     0.675493     0.772834  0.747396  0.724285  0.796723       8      MRM
2     0.633127     0.743852  0.710935  0.688628  0.732537       8     Mean
3     0.430421     0.685597  0.576665  0.558327  0.588858       8    Rmssd
4     0.577817     0.671838  0.636629  0.624945  0.652066       8  MissVal
5     0.470138     0.520199  0.494885  0.497872  0.505699       8     Rand
0     0.744211     0.823977  0.778372  0.787783  0.849315      12      Sig
1     0.672321     0.826884  0.763375  0.756750  0.812551      12      MRM
2     0.623856     0.700291  0.633579  0.665608  0.719829      12     Mean
3     0.405223     0.711921  0.538847  0.572755  0.581289      12    Rmssd
4     0.592623     0.751062  0.664152  0.679169  0.709493      12  MissVal
5     0.543888     0.463655  0.450788  0.496762  0.490970      12     Rand
0     0.775064     0.794737  0.821246  0.783939  0.855578      20      Sig
1     0.734351     0.745820  0.778527  0.739525  0.809667      20      MRM
2     0.571756     0.759133  0.742810  0.656285  0.724444      20     Mean
3     0.649109     0.433437  0.582287  0.551816  0.568039      20    Rmssd
4     0.540712     0.820743  0.785873  0.667039  0.741546      20  MissVal
5     0.550127     0.448916  0.546867  0.506564  0.498109      20     Rand
0     0.813607     0.824252  0.891650  0.817440  0.889695      50      Sig
1     0.671687     0.789921  0.850381  0.714253  0.826069      50      MRM
2     0.677002     0.769764  0.839411  0.710398  0.772619      50     Mean
3     0.410524     0.658583  0.681270  0.499830  0.563487      50    Rmssd
4     0.630758     0.773543  0.831970  0.682164  0.778421      50  MissVal
5     0.609142     0.384882  0.638997  0.525116  0.496973      50     Rand

Answer 1

您必须先融合数据，以便列名成为另一列中的值。

print(df.head(10))

   Sensitivity  Specificity  Accuracy       PPV       AUC  k_week    model
0     0.609948     0.867516  0.717291  0.776004  0.813680       4      Sig
1     0.586824     0.851647  0.685525  0.757557  0.784805       4      MRM
2     0.581152     0.842510  0.670357  0.749651  0.737269       4     Mean
3     0.473386     0.601827  0.395841  0.556193  0.542679       4    Rmssd
4     0.511344     0.717240  0.499148  0.644086  0.626719       4  MissVal
5     0.454625     0.523924  0.356061  0.508448  0.501661       4     Rand
0     0.792619     0.764423  0.755839  0.777934  0.835871       6      Sig
1     0.722077     0.758113  0.733090  0.740845  0.803949       6      MRM
2     0.633899     0.735877  0.688298  0.687011  0.736115       6     Mean
3     0.437296     0.613582  0.510095  0.529108  0.544014       6    Rmssd

df = df.melt(id_vars=['k_week', 'model'])
print(df.head(10))

   k_week    model     variable     value
0       4      Sig  Sensitivity  0.609948
1       4      MRM  Sensitivity  0.586824
2       4     Mean  Sensitivity  0.581152
3       4    Rmssd  Sensitivity  0.473386
4       4  MissVal  Sensitivity  0.511344
5       4     Rand  Sensitivity  0.454625
6       6      Sig  Sensitivity  0.792619
7       6      MRM  Sensitivity  0.722077
8       6     Mean  Sensitivity  0.633899
9       6    Rmssd  Sensitivity  0.437296

grid = sns.FacetGrid(df, row='k_week', col='variable')
grid.map_dataframe(sns.boxplot, x='model', y='value')