python - 如何使用汇总统计注释seaborn PairGrid 对角线

Question

我尝试通过以下代码将有关每个参数的统计数据放置在绘图的对角线上:

import matplotlib.pyplot as plt
import seaborn as sns
iris = sns.load_dataset('iris')

def summary(x, **kwargs):
    x = pd.Series(x)
    
    label = x.describe()[['mean', 'std', 'min', '50%', 'max']]
       
    label = label.round()
    ax = plt.gca()
    ax.set_axis_off()
    
    ax.annotate(pd.DataFrame(label),
               xy = (0.1, 0.2), size = 20, xycoords = ax.transAxes)

grd = sns.PairGrid(data=iris, size = 4)
grd = grd.map_upper(plt.scatter, color = 'k')

grd = grd.map_lower(sns.kdeplot, cmap = 'PRGn_r')
grd = grd.map_upper(sns.kdeplot, cmap = 'PRGn_r')
grd = grd.map_lower(plt.scatter, color = 'k')

grd = grd.map_diag(summary);

但由于某种原因，我遇到了 ValueError: DataFrame 的真值不明确。在尝试执行 summary 函数时使用 a.empty、a.bool()、a.item()、a.any() 或 a.all().精确 - 将数据集 pd.DataFrame(label) 放入 ax.annotate 中。

回溯

-------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-26-88744ab705f0> in <module>
      6 grd = grd.map_lower(plt.scatter, color = 'k')
      7 
----> 8 grd = grd.map_diag(summary)
      9 plt.show()

e:\Anaconda3\lib\site-packages\seaborn\axisgrid.py in map_diag(self, func, **kwargs)
   1448 
   1449         if "hue" not in signature(func).parameters:
-> 1450             return self._map_diag_iter_hue(func, **kwargs)
   1451 
   1452         # Loop over diagonal variables and axes, making one plot in each

e:\Anaconda3\lib\site-packages\seaborn\axisgrid.py in _map_diag_iter_hue(self, func, **kwargs)
   1515                     func(x=data_k, label=label_k, color=color, **plot_kwargs)
   1516                 else:
-> 1517                     func(data_k, label=label_k, color=color, **plot_kwargs)
   1518 
   1519         self._add_axis_labels()

<ipython-input-25-d3b35ec81d97> in summary(x, **kwargs)
      8     ax.set_axis_off()
      9 
---> 10     ax.annotate(pd.DataFrame(label),
     11                xy = (0.1, 0.2), size = 20, xycoords = ax.transAxes)

e:\Anaconda3\lib\site-packages\matplotlib\_api\deprecation.py in wrapper(*args, **kwargs)
    333                 f"for the old name will be dropped %(removal)s.")
    334             kwargs[new] = kwargs.pop(old)
--> 335         return func(*args, **kwargs)
    336 
    337     # wrapper() must keep the same documented signature as func(): if we

e:\Anaconda3\lib\site-packages\matplotlib\axes\_axes.py in annotate(self, text, xy, *args, **kwargs)
    650     @docstring.dedent_interpd
    651     def annotate(self, text, xy, *args, **kwargs):
--> 652         a = mtext.Annotation(text, xy, *args, **kwargs)
    653         a.set_transform(mtransforms.IdentityTransform())
    654         if 'clip_on' in kwargs:

e:\Anaconda3\lib\site-packages\matplotlib\text.py in __init__(self, text, xy, xytext, xycoords, textcoords, arrowprops, annotation_clip, **kwargs)
   1804 
   1805         # Must come last, as some kwargs may be propagated to arrow_patch.
-> 1806         Text.__init__(self, x, y, text, **kwargs)
   1807 
   1808     def contains(self, event):

e:\Anaconda3\lib\site-packages\matplotlib\text.py in __init__(self, x, y, text, color, verticalalignment, horizontalalignment, multialignment, fontproperties, rotation, linespacing, rotation_mode, usetex, wrap, transform_rotates_text, **kwargs)
    152         self._x, self._y = x, y
    153         self._text = ''
--> 154         self.set_text(text)
    155         self.set_color(
    156             color if color is not None else mpl.rcParams["text.color"])

e:\Anaconda3\lib\site-packages\matplotlib\text.py in set_text(self, s)
   1213         if s is None:
   1214             s = ''
-> 1215         if s != self._text:
   1216             self._text = str(s)
   1217             self.stale = True

e:\Anaconda3\lib\site-packages\pandas\core\generic.py in __nonzero__(self)
   1532     @final
   1533     def __nonzero__(self):
-> 1534         raise ValueError(
   1535             f"The truth value of a {type(self).__name__} is ambiguous. "
   1536             "Use a.empty, a.bool(), a.item(), a.any() or a.all()."

ValueError: The truth value of a DataFrame is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

Answer 1

• .annotate需要一个 string 而不是 DataFrame 或 Series。使用 .to_string() 将系列、标签转换为字符串
• x 已经是 pandas.Series，因此不需要 x = pd.Series(x)。
• 可重现并使用 seaborn 0.11.2、matplotlib 3.4.2 和 pandas 1.3.1 进行测试

import seaborn as sns
import matplotlib.pyplot as plt

# change your function; ax.annotate expects a string
def summary(x, **kwargs):
    
    label = x.describe()[['mean', 'std', 'min', '50%', 'max']].round()
       
    ax = plt.gca()
    ax.set_axis_off()
    
    # use .to_string()   
    ax.annotate(label.to_string(), xy=(0.1, 0.2), size=20, xycoords=ax.transAxes)