Display both absolute frequencies and percentages

CHANGELOG.rst

@@ -1,6 +1,7 @@
 In development
 --------------
-
+- Added option to display both the absolute frequency and the percentage of
+  the total for each intersection and category. (:issue:`89`)
 - Improved efficiency where there are many categories, but valid combinations
   are sparse, if `sort_by='degree'`. (:issue:`82`)
 - Permit truthy (not necessarily bool) values in index. (:issue:`74`)

examples/plot_generated.py

@@ -23,3 +23,7 @@
 plot(example, show_counts='%d')
 plt.suptitle('With counts shown')
 plt.show()
+
+plot(example, show_counts='%d', show_percentages=True)
+plt.suptitle('With counts and % shown')
+plt.show()

examples/plot_vertical.py

@@ -17,3 +17,7 @@
 plot(example, orientation='vertical', show_counts='%d')
 plt.suptitle('A vertical plot with counts shown')
 plt.show()
+
+plot(example, orientation='vertical', show_counts='%d', show_percentages=True)
+plt.suptitle('With counts and percentages shown')
+plt.show()

upsetplot/__init__.py

@@ -1,4 +1,4 @@
-__version__ = '0.4-dev'
+__version__ = '0.4.dev1'
 
 import os
 

upsetplot/plotting.py

@@ -272,6 +272,10 @@ class UpSet:
         Whether to label the intersection size bars with the cardinality
         of the intersection. When a string, this formats the number.
         For example, '%d' is equivalent to True.
+    show_percentages : bool, default=False
+        Whether to label the intersection size bars with the percentage
+        of the intersection relative to the total dataset.
+        This may be applied with or without show_counts.
     sort_sets_by
         .. deprecated: 0.3
             Replaced by sort_categories_by, this parameter will be removed in
@@ -285,7 +289,8 @@ def __init__(self, data, orientation='horizontal', sort_by='degree',
                  facecolor='black',
                  with_lines=True, element_size=32,
                  intersection_plot_elements=6, totals_plot_elements=2,
-                 show_counts='', sort_sets_by='deprecated'):
+                 show_counts='', show_percentages=False,
+                 sort_sets_by='deprecated'):
 
         self._horizontal = orientation == 'horizontal'
         self._reorient = _identity if self._horizontal else _transpose
@@ -299,6 +304,7 @@ def __init__(self, data, orientation='horizontal', sort_by='degree',
         if not intersection_plot_elements:
             self._subset_plots.pop()
         self._show_counts = show_counts
+        self._show_percentages = show_percentages
 
         if sort_sets_by != 'deprecated':
             sort_categories_by = sort_sets_by
@@ -429,7 +435,6 @@ def make_grid(self, fig=None):
                                       self._totals_plot_elements),
                       hspace=1)
         if self._horizontal:
-            print(n_cats, n_inters, self._totals_plot_elements)
             out = {'matrix': gridspec[-n_cats:, -n_inters:],
                    'shading': gridspec[-n_cats:, :],
                    'totals': gridspec[-n_cats:, :self._totals_plot_elements],
@@ -446,8 +451,8 @@ def make_grid(self, fig=None):
             cumsizes = np.cumsum(sizes)
             for start, stop, plot in zip(np.hstack([[0], cumsizes]), cumsizes,
                                          self._subset_plots):
-                out[plot['id']] = gridspec[-n_inters:,
-                                           start + n_cats:stop + n_cats]
+                out[plot['id']] = \
+                    gridspec[-n_inters:, start + n_cats:stop + n_cats]
         return out
 
     def plot_matrix(self, ax):
@@ -505,31 +510,60 @@ def plot_intersections(self, ax):
         ax.set_ylabel('Intersection size')
 
     def _label_sizes(self, ax, rects, where):
-        if not self._show_counts:
+        if not self._show_counts and not self._show_percentages:
             return
-        fmt = '%d' if self._show_counts is True else self._show_counts
+        if self._show_counts is True:
+            count_fmt = "%d"
+        else:
+            count_fmt = self._show_counts
+        if self._show_percentages is True:
+            pct_fmt = "%.1f%%"
+        else:
+            pct_fmt = self._show_percentages
+
+        total = sum(self.totals)
+        if count_fmt and pct_fmt:
+            if where == 'top':
+                fmt = '%s\n(%s)' % (count_fmt, pct_fmt)
+            else:
+                fmt = '%s (%s)' % (count_fmt, pct_fmt)
+
+            def make_args(val):
+                return val, 100 * val / total
+        elif count_fmt:
+            fmt = count_fmt
+
+            def make_args(val):
+                return val,
+        else:
+            fmt = pct_fmt
+
+            def make_args(val):
+                return 100 * val / total,
+
         if where == 'right':
             margin = 0.01 * abs(np.diff(ax.get_xlim()))
             for rect in rects:
                 width = rect.get_width()
                 ax.text(width + margin,
                         rect.get_y() + rect.get_height() * .5,
-                        fmt % width,
+                        fmt % make_args(width),
                         ha='left', va='center')
         elif where == 'left':
             margin = 0.01 * abs(np.diff(ax.get_xlim()))
             for rect in rects:
                 width = rect.get_width()
                 ax.text(width + margin,
                         rect.get_y() + rect.get_height() * .5,
-                        fmt % width,
+                        fmt % make_args(width),
                         ha='right', va='center')
         elif where == 'top':
             margin = 0.01 * abs(np.diff(ax.get_ylim()))
             for rect in rects:
                 height = rect.get_height()
                 ax.text(rect.get_x() + rect.get_width() * .5,
-                        height + margin, fmt % height,
+                        height + margin,
+                        fmt % make_args(height),
                         ha='center', va='bottom')
         else:
             raise NotImplementedError('unhandled where: %r' % where)

upsetplot/tests/test_upsetplot.py

@@ -9,6 +9,7 @@
 import numpy as np
 import matplotlib.figure
 import matplotlib.pyplot as plt
+from matplotlib.text import Text
 
 from upsetplot import plot
 from upsetplot import UpSet
@@ -23,6 +24,11 @@ def is_ascending(seq):
     return sorted(seq) == list(seq)
 
 
+def get_all_texts(mpl_artist):
+    out = [text.get_text() for text in mpl_artist.findobj(Text)]
+    return [text for text in out if text]
+
+
 @pytest.mark.parametrize('x', [
     generate_counts(),
     generate_counts().iloc[1:-2],
@@ -389,22 +395,43 @@ def _count_descendants(el):
 @pytest.mark.parametrize('orientation', ['horizontal', 'vertical'])
 def test_show_counts(orientation):
     fig = matplotlib.figure.Figure()
-    X = generate_counts(n_samples=100)
-    plot(X, fig)
+    X = generate_counts(n_samples=10000)
+    plot(X, fig, orientation=orientation)
     n_artists_no_sizes = _count_descendants(fig)
 
     fig = matplotlib.figure.Figure()
-    plot(X, fig, show_counts=True)
+    plot(X, fig, orientation=orientation, show_counts=True)
     n_artists_yes_sizes = _count_descendants(fig)
     assert n_artists_yes_sizes - n_artists_no_sizes > 6
+    assert '9547' in get_all_texts(fig)  # set size
+    assert '283' in get_all_texts(fig)   # intersection size
+
+    fig = matplotlib.figure.Figure()
+    plot(X, fig, orientation=orientation, show_counts='%0.2g')
+    assert n_artists_yes_sizes == _count_descendants(fig)
+    assert '9.5e+03' in get_all_texts(fig)
+    assert '2.8e+02' in get_all_texts(fig)
 
     fig = matplotlib.figure.Figure()
-    plot(X, fig, show_counts='%0.2g')
+    plot(X, fig, orientation=orientation, show_percentages=True)
     assert n_artists_yes_sizes == _count_descendants(fig)
+    assert '47.1%' in get_all_texts(fig)
+    assert '1.4%' in get_all_texts(fig)
+
+    fig = matplotlib.figure.Figure()
+    plot(X, fig, orientation=orientation, show_counts=True,
+         show_percentages=True)
+    assert n_artists_yes_sizes == _count_descendants(fig)
+    if orientation == 'vertical':
+        assert '9547\n(47.1%)' in get_all_texts(fig)
+        assert '283 (1.4%)' in get_all_texts(fig)
+    else:
+        assert '9547 (47.1%)' in get_all_texts(fig)
+        assert '283\n(1.4%)' in get_all_texts(fig)
 
     with pytest.raises(ValueError):
         fig = matplotlib.figure.Figure()
-        plot(X, fig, show_counts='%0.2h')
+        plot(X, fig, orientation=orientation, show_counts='%0.2h')
 
 
 def test_add_catplot():