-
Notifications
You must be signed in to change notification settings - Fork 1.9k
/
axisgrid.py
2350 lines (1875 loc) · 83.7 KB
/
axisgrid.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
from itertools import product
import warnings
from textwrap import dedent
from distutils.version import LooseVersion
import numpy as np
import pandas as pd
from scipy import stats
import matplotlib as mpl
import matplotlib.pyplot as plt
from . import utils
from .palettes import color_palette, blend_palette
from .distributions import distplot, kdeplot, _freedman_diaconis_bins
__all__ = ["FacetGrid", "PairGrid", "JointGrid", "pairplot", "jointplot"]
class Grid(object):
"""Base class for grids of subplots."""
_margin_titles = False
_legend_out = True
def set(self, **kwargs):
"""Set attributes on each subplot Axes."""
for ax in self.axes.flat:
ax.set(**kwargs)
return self
def savefig(self, *args, **kwargs):
"""Save the figure."""
kwargs = kwargs.copy()
kwargs.setdefault("bbox_inches", "tight")
self.fig.savefig(*args, **kwargs)
def add_legend(self, legend_data=None, title=None, label_order=None,
**kwargs):
"""Draw a legend, maybe placing it outside axes and resizing the figure.
Parameters
----------
legend_data : dict, optional
Dictionary mapping label names (or two-element tuples where the
second element is a label name) to matplotlib artist handles. The
default reads from ``self._legend_data``.
title : string, optional
Title for the legend. The default reads from ``self._hue_var``.
label_order : list of labels, optional
The order that the legend entries should appear in. The default
reads from ``self.hue_names``.
kwargs : key, value pairings
Other keyword arguments are passed to the underlying legend methods
on the Figure or Axes object.
Returns
-------
self : Grid instance
Returns self for easy chaining.
"""
# Find the data for the legend
if legend_data is None:
legend_data = self._legend_data
if label_order is None:
if self.hue_names is None:
label_order = list(legend_data.keys())
else:
label_order = list(map(utils.to_utf8, self.hue_names))
blank_handle = mpl.patches.Patch(alpha=0, linewidth=0)
handles = [legend_data.get(l, blank_handle) for l in label_order]
title = self._hue_var if title is None else title
if LooseVersion(mpl.__version__) < LooseVersion("3.0"):
try:
title_size = mpl.rcParams["axes.labelsize"] * .85
except TypeError: # labelsize is something like "large"
title_size = mpl.rcParams["axes.labelsize"]
else:
title_size = mpl.rcParams["legend.title_fontsize"]
# Unpack nested labels from a hierarchical legend
labels = []
for entry in label_order:
if isinstance(entry, tuple):
_, label = entry
else:
label = entry
labels.append(label)
# Set default legend kwargs
kwargs.setdefault("scatterpoints", 1)
if self._legend_out:
kwargs.setdefault("frameon", False)
kwargs.setdefault("loc", "center right")
# Draw a full-figure legend outside the grid
figlegend = self.fig.legend(handles, labels, **kwargs)
self._legend = figlegend
figlegend.set_title(title, prop={"size": title_size})
# Draw the plot to set the bounding boxes correctly
if hasattr(self.fig.canvas, "get_renderer"):
self.fig.draw(self.fig.canvas.get_renderer())
# Calculate and set the new width of the figure so the legend fits
legend_width = figlegend.get_window_extent().width / self.fig.dpi
fig_width, fig_height = self.fig.get_size_inches()
self.fig.set_size_inches(fig_width + legend_width, fig_height)
# Draw the plot again to get the new transformations
if hasattr(self.fig.canvas, "get_renderer"):
self.fig.draw(self.fig.canvas.get_renderer())
# Now calculate how much space we need on the right side
legend_width = figlegend.get_window_extent().width / self.fig.dpi
space_needed = legend_width / (fig_width + legend_width)
margin = .04 if self._margin_titles else .01
self._space_needed = margin + space_needed
right = 1 - self._space_needed
# Place the subplot axes to give space for the legend
self.fig.subplots_adjust(right=right)
else:
# Draw a legend in the first axis
ax = self.axes.flat[0]
kwargs.setdefault("loc", "best")
leg = ax.legend(handles, labels, **kwargs)
leg.set_title(title, prop={"size": title_size})
self._legend = leg
return self
def _clean_axis(self, ax):
"""Turn off axis labels and legend."""
ax.set_xlabel("")
ax.set_ylabel("")
ax.legend_ = None
return self
def _update_legend_data(self, ax):
"""Extract the legend data from an axes object and save it."""
handles, labels = ax.get_legend_handles_labels()
data = {l: h for h, l in zip(handles, labels)}
self._legend_data.update(data)
def _get_palette(self, data, hue, hue_order, palette):
"""Get a list of colors for the hue variable."""
if hue is None:
palette = color_palette(n_colors=1)
else:
hue_names = utils.categorical_order(data[hue], hue_order)
n_colors = len(hue_names)
# By default use either the current color palette or HUSL
if palette is None:
current_palette = utils.get_color_cycle()
if n_colors > len(current_palette):
colors = color_palette("husl", n_colors)
else:
colors = color_palette(n_colors=n_colors)
# Allow for palette to map from hue variable names
elif isinstance(palette, dict):
color_names = [palette[h] for h in hue_names]
colors = color_palette(color_names, n_colors)
# Otherwise act as if we just got a list of colors
else:
colors = color_palette(palette, n_colors)
palette = color_palette(colors, n_colors)
return palette
_facet_docs = dict(
data=dedent("""\
data : DataFrame
Tidy ("long-form") dataframe where each column is a variable and each
row is an observation.\
"""),
col_wrap=dedent("""\
col_wrap : int, optional
"Wrap" the column variable at this width, so that the column facets
span multiple rows. Incompatible with a ``row`` facet.\
"""),
share_xy=dedent("""\
share{x,y} : bool, 'col', or 'row' optional
If true, the facets will share y axes across columns and/or x axes
across rows.\
"""),
height=dedent("""\
height : scalar, optional
Height (in inches) of each facet. See also: ``aspect``.\
"""),
aspect=dedent("""\
aspect : scalar, optional
Aspect ratio of each facet, so that ``aspect * height`` gives the width
of each facet in inches.\
"""),
palette=dedent("""\
palette : palette name, list, or dict, optional
Colors to use for the different levels of the ``hue`` variable. Should
be something that can be interpreted by :func:`color_palette`, or a
dictionary mapping hue levels to matplotlib colors.\
"""),
legend_out=dedent("""\
legend_out : bool, optional
If ``True``, the figure size will be extended, and the legend will be
drawn outside the plot on the center right.\
"""),
margin_titles=dedent("""\
margin_titles : bool, optional
If ``True``, the titles for the row variable are drawn to the right of
the last column. This option is experimental and may not work in all
cases.\
"""),
)
class FacetGrid(Grid):
"""Multi-plot grid for plotting conditional relationships."""
def __init__(self, data, row=None, col=None, hue=None, col_wrap=None,
sharex=True, sharey=True, height=3, aspect=1, palette=None,
row_order=None, col_order=None, hue_order=None, hue_kws=None,
dropna=True, legend_out=True, despine=True,
margin_titles=False, xlim=None, ylim=None, subplot_kws=None,
gridspec_kws=None, size=None):
# Handle deprecations
if size is not None:
height = size
msg = ("The `size` parameter has been renamed to `height`; "
"please update your code.")
warnings.warn(msg, UserWarning)
# Determine the hue facet layer information
hue_var = hue
if hue is None:
hue_names = None
else:
hue_names = utils.categorical_order(data[hue], hue_order)
colors = self._get_palette(data, hue, hue_order, palette)
# Set up the lists of names for the row and column facet variables
if row is None:
row_names = []
else:
row_names = utils.categorical_order(data[row], row_order)
if col is None:
col_names = []
else:
col_names = utils.categorical_order(data[col], col_order)
# Additional dict of kwarg -> list of values for mapping the hue var
hue_kws = hue_kws if hue_kws is not None else {}
# Make a boolean mask that is True anywhere there is an NA
# value in one of the faceting variables, but only if dropna is True
none_na = np.zeros(len(data), np.bool)
if dropna:
row_na = none_na if row is None else data[row].isnull()
col_na = none_na if col is None else data[col].isnull()
hue_na = none_na if hue is None else data[hue].isnull()
not_na = ~(row_na | col_na | hue_na)
else:
not_na = ~none_na
# Compute the grid shape
ncol = 1 if col is None else len(col_names)
nrow = 1 if row is None else len(row_names)
self._n_facets = ncol * nrow
self._col_wrap = col_wrap
if col_wrap is not None:
if row is not None:
err = "Cannot use `row` and `col_wrap` together."
raise ValueError(err)
ncol = col_wrap
nrow = int(np.ceil(len(col_names) / col_wrap))
self._ncol = ncol
self._nrow = nrow
# Calculate the base figure size
# This can get stretched later by a legend
# TODO this doesn't account for axis labels
figsize = (ncol * height * aspect, nrow * height)
# Validate some inputs
if col_wrap is not None:
margin_titles = False
# Build the subplot keyword dictionary
subplot_kws = {} if subplot_kws is None else subplot_kws.copy()
gridspec_kws = {} if gridspec_kws is None else gridspec_kws.copy()
if xlim is not None:
subplot_kws["xlim"] = xlim
if ylim is not None:
subplot_kws["ylim"] = ylim
# Initialize the subplot grid
if col_wrap is None:
kwargs = dict(figsize=figsize, squeeze=False,
sharex=sharex, sharey=sharey,
subplot_kw=subplot_kws,
gridspec_kw=gridspec_kws)
fig, axes = plt.subplots(nrow, ncol, **kwargs)
self.axes = axes
else:
# If wrapping the col variable we need to make the grid ourselves
if gridspec_kws:
warnings.warn("`gridspec_kws` ignored when using `col_wrap`")
n_axes = len(col_names)
fig = plt.figure(figsize=figsize)
axes = np.empty(n_axes, object)
axes[0] = fig.add_subplot(nrow, ncol, 1, **subplot_kws)
if sharex:
subplot_kws["sharex"] = axes[0]
if sharey:
subplot_kws["sharey"] = axes[0]
for i in range(1, n_axes):
axes[i] = fig.add_subplot(nrow, ncol, i + 1, **subplot_kws)
self.axes = axes
# Now we turn off labels on the inner axes
if sharex:
for ax in self._not_bottom_axes:
for label in ax.get_xticklabels():
label.set_visible(False)
ax.xaxis.offsetText.set_visible(False)
if sharey:
for ax in self._not_left_axes:
for label in ax.get_yticklabels():
label.set_visible(False)
ax.yaxis.offsetText.set_visible(False)
# Set up the class attributes
# ---------------------------
# First the public API
self.data = data
self.fig = fig
self.axes = axes
self.row_names = row_names
self.col_names = col_names
self.hue_names = hue_names
self.hue_kws = hue_kws
# Next the private variables
self._nrow = nrow
self._row_var = row
self._ncol = ncol
self._col_var = col
self._margin_titles = margin_titles
self._col_wrap = col_wrap
self._hue_var = hue_var
self._colors = colors
self._legend_out = legend_out
self._legend = None
self._legend_data = {}
self._x_var = None
self._y_var = None
self._dropna = dropna
self._not_na = not_na
# Make the axes look good
fig.tight_layout()
if despine:
self.despine()
__init__.__doc__ = dedent("""\
Initialize the matplotlib figure and FacetGrid object.
This class maps a dataset onto multiple axes arrayed in a grid of rows
and columns that correspond to *levels* of variables in the dataset.
The plots it produces are often called "lattice", "trellis", or
"small-multiple" graphics.
It can also represent levels of a third variable with the ``hue``
parameter, which plots different subsets of data in different colors.
This uses color to resolve elements on a third dimension, but only
draws subsets on top of each other and will not tailor the ``hue``
parameter for the specific visualization the way that axes-level
functions that accept ``hue`` will.
When using seaborn functions that infer semantic mappings from a
dataset, care must be taken to synchronize those mappings across
facets (e.g., by defing the ``hue`` mapping with a palette dict or
setting the data type of the variables to ``category``). In most cases,
it will be better to use a figure-level function (e.g. :func:`relplot`
or :func:`catplot`) than to use :class:`FacetGrid` directly.
The basic workflow is to initialize the :class:`FacetGrid` object with
the dataset and the variables that are used to structure the grid. Then
one or more plotting functions can be applied to each subset by calling
:meth:`FacetGrid.map` or :meth:`FacetGrid.map_dataframe`. Finally, the
plot can be tweaked with other methods to do things like change the
axis labels, use different ticks, or add a legend. See the detailed
code examples below for more information.
See the :ref:`tutorial <grid_tutorial>` for more information.
Parameters
----------
{data}
row, col, hue : strings
Variables that define subsets of the data, which will be drawn on
separate facets in the grid. See the ``*_order`` parameters to
control the order of levels of this variable.
{col_wrap}
{share_xy}
{height}
{aspect}
{palette}
{{row,col,hue}}_order : lists, optional
Order for the levels of the faceting variables. By default, this
will be the order that the levels appear in ``data`` or, if the
variables are pandas categoricals, the category order.
hue_kws : dictionary of param -> list of values mapping
Other keyword arguments to insert into the plotting call to let
other plot attributes vary across levels of the hue variable (e.g.
the markers in a scatterplot).
{legend_out}
despine : boolean, optional
Remove the top and right spines from the plots.
{margin_titles}
{{x, y}}lim: tuples, optional
Limits for each of the axes on each facet (only relevant when
share{{x, y}} is True).
subplot_kws : dict, optional
Dictionary of keyword arguments passed to matplotlib subplot(s)
methods.
gridspec_kws : dict, optional
Dictionary of keyword arguments passed to matplotlib's ``gridspec``
module (via ``plt.subplots``). Ignored if ``col_wrap`` is not
``None``.
See Also
--------
PairGrid : Subplot grid for plotting pairwise relationships.
relplot : Combine a relational plot and a :class:`FacetGrid`.
catplot : Combine a categorical plot and a :class:`FacetGrid`.
lmplot : Combine a regression plot and a :class:`FacetGrid`.
Examples
--------
Initialize a 2x2 grid of facets using the tips dataset:
.. plot::
:context: close-figs
>>> import seaborn as sns; sns.set(style="ticks", color_codes=True)
>>> tips = sns.load_dataset("tips")
>>> g = sns.FacetGrid(tips, col="time", row="smoker")
Draw a univariate plot on each facet:
.. plot::
:context: close-figs
>>> import matplotlib.pyplot as plt
>>> g = sns.FacetGrid(tips, col="time", row="smoker")
>>> g = g.map(plt.hist, "total_bill")
(Note that it's not necessary to re-catch the returned variable; it's
the same object, but doing so in the examples makes dealing with the
doctests somewhat less annoying).
Pass additional keyword arguments to the mapped function:
.. plot::
:context: close-figs
>>> import numpy as np
>>> bins = np.arange(0, 65, 5)
>>> g = sns.FacetGrid(tips, col="time", row="smoker")
>>> g = g.map(plt.hist, "total_bill", bins=bins, color="r")
Plot a bivariate function on each facet:
.. plot::
:context: close-figs
>>> g = sns.FacetGrid(tips, col="time", row="smoker")
>>> g = g.map(plt.scatter, "total_bill", "tip", edgecolor="w")
Assign one of the variables to the color of the plot elements:
.. plot::
:context: close-figs
>>> g = sns.FacetGrid(tips, col="time", hue="smoker")
>>> g = (g.map(plt.scatter, "total_bill", "tip", edgecolor="w")
... .add_legend())
Change the height and aspect ratio of each facet:
.. plot::
:context: close-figs
>>> g = sns.FacetGrid(tips, col="day", height=4, aspect=.5)
>>> g = g.map(plt.hist, "total_bill", bins=bins)
Specify the order for plot elements:
.. plot::
:context: close-figs
>>> g = sns.FacetGrid(tips, col="smoker", col_order=["Yes", "No"])
>>> g = g.map(plt.hist, "total_bill", bins=bins, color="m")
Use a different color palette:
.. plot::
:context: close-figs
>>> kws = dict(s=50, linewidth=.5, edgecolor="w")
>>> g = sns.FacetGrid(tips, col="sex", hue="time", palette="Set1",
... hue_order=["Dinner", "Lunch"])
>>> g = (g.map(plt.scatter, "total_bill", "tip", **kws)
... .add_legend())
Use a dictionary mapping hue levels to colors:
.. plot::
:context: close-figs
>>> pal = dict(Lunch="seagreen", Dinner="gray")
>>> g = sns.FacetGrid(tips, col="sex", hue="time", palette=pal,
... hue_order=["Dinner", "Lunch"])
>>> g = (g.map(plt.scatter, "total_bill", "tip", **kws)
... .add_legend())
Additionally use a different marker for the hue levels:
.. plot::
:context: close-figs
>>> g = sns.FacetGrid(tips, col="sex", hue="time", palette=pal,
... hue_order=["Dinner", "Lunch"],
... hue_kws=dict(marker=["^", "v"]))
>>> g = (g.map(plt.scatter, "total_bill", "tip", **kws)
... .add_legend())
"Wrap" a column variable with many levels into the rows:
.. plot::
:context: close-figs
>>> att = sns.load_dataset("attention")
>>> g = sns.FacetGrid(att, col="subject", col_wrap=5, height=1.5)
>>> g = g.map(plt.plot, "solutions", "score", marker=".")
Define a custom bivariate function to map onto the grid:
.. plot::
:context: close-figs
>>> from scipy import stats
>>> def qqplot(x, y, **kwargs):
... _, xr = stats.probplot(x, fit=False)
... _, yr = stats.probplot(y, fit=False)
... sns.scatterplot(xr, yr, **kwargs)
>>> g = sns.FacetGrid(tips, col="smoker", hue="sex")
>>> g = (g.map(qqplot, "total_bill", "tip", **kws)
... .add_legend())
Define a custom function that uses a ``DataFrame`` object and accepts
column names as positional variables:
.. plot::
:context: close-figs
>>> import pandas as pd
>>> df = pd.DataFrame(
... data=np.random.randn(90, 4),
... columns=pd.Series(list("ABCD"), name="walk"),
... index=pd.date_range("2015-01-01", "2015-03-31",
... name="date"))
>>> df = df.cumsum(axis=0).stack().reset_index(name="val")
>>> def dateplot(x, y, **kwargs):
... ax = plt.gca()
... data = kwargs.pop("data")
... data.plot(x=x, y=y, ax=ax, grid=False, **kwargs)
>>> g = sns.FacetGrid(df, col="walk", col_wrap=2, height=3.5)
>>> g = g.map_dataframe(dateplot, "date", "val")
Use different axes labels after plotting:
.. plot::
:context: close-figs
>>> g = sns.FacetGrid(tips, col="smoker", row="sex")
>>> g = (g.map(plt.scatter, "total_bill", "tip", color="g", **kws)
... .set_axis_labels("Total bill (US Dollars)", "Tip"))
Set other attributes that are shared across the facetes:
.. plot::
:context: close-figs
>>> g = sns.FacetGrid(tips, col="smoker", row="sex")
>>> g = (g.map(plt.scatter, "total_bill", "tip", color="r", **kws)
... .set(xlim=(0, 60), ylim=(0, 12),
... xticks=[10, 30, 50], yticks=[2, 6, 10]))
Use a different template for the facet titles:
.. plot::
:context: close-figs
>>> g = sns.FacetGrid(tips, col="size", col_wrap=3)
>>> g = (g.map(plt.hist, "tip", bins=np.arange(0, 13), color="c")
... .set_titles("{{col_name}} diners"))
Tighten the facets:
.. plot::
:context: close-figs
>>> g = sns.FacetGrid(tips, col="smoker", row="sex",
... margin_titles=True)
>>> g = (g.map(plt.scatter, "total_bill", "tip", color="m", **kws)
... .set(xlim=(0, 60), ylim=(0, 12),
... xticks=[10, 30, 50], yticks=[2, 6, 10])
... .fig.subplots_adjust(wspace=.05, hspace=.05))
""").format(**_facet_docs)
def facet_data(self):
"""Generator for name indices and data subsets for each facet.
Yields
------
(i, j, k), data_ijk : tuple of ints, DataFrame
The ints provide an index into the {row, col, hue}_names attribute,
and the dataframe contains a subset of the full data corresponding
to each facet. The generator yields subsets that correspond with
the self.axes.flat iterator, or self.axes[i, j] when `col_wrap`
is None.
"""
data = self.data
# Construct masks for the row variable
if self.row_names:
row_masks = [data[self._row_var] == n for n in self.row_names]
else:
row_masks = [np.repeat(True, len(self.data))]
# Construct masks for the column variable
if self.col_names:
col_masks = [data[self._col_var] == n for n in self.col_names]
else:
col_masks = [np.repeat(True, len(self.data))]
# Construct masks for the hue variable
if self.hue_names:
hue_masks = [data[self._hue_var] == n for n in self.hue_names]
else:
hue_masks = [np.repeat(True, len(self.data))]
# Here is the main generator loop
for (i, row), (j, col), (k, hue) in product(enumerate(row_masks),
enumerate(col_masks),
enumerate(hue_masks)):
data_ijk = data[row & col & hue & self._not_na]
yield (i, j, k), data_ijk
def map(self, func, *args, **kwargs):
"""Apply a plotting function to each facet's subset of the data.
Parameters
----------
func : callable
A plotting function that takes data and keyword arguments. It
must plot to the currently active matplotlib Axes and take a
`color` keyword argument. If faceting on the `hue` dimension,
it must also take a `label` keyword argument.
args : strings
Column names in self.data that identify variables with data to
plot. The data for each variable is passed to `func` in the
order the variables are specified in the call.
kwargs : keyword arguments
All keyword arguments are passed to the plotting function.
Returns
-------
self : object
Returns self.
"""
# If color was a keyword argument, grab it here
kw_color = kwargs.pop("color", None)
if hasattr(func, "__module__"):
func_module = str(func.__module__)
else:
func_module = ""
# Check for categorical plots without order information
if func_module == "seaborn.categorical":
if "order" not in kwargs:
warning = ("Using the {} function without specifying "
"`order` is likely to produce an incorrect "
"plot.".format(func.__name__))
warnings.warn(warning)
if len(args) == 3 and "hue_order" not in kwargs:
warning = ("Using the {} function without specifying "
"`hue_order` is likely to produce an incorrect "
"plot.".format(func.__name__))
warnings.warn(warning)
# Iterate over the data subsets
for (row_i, col_j, hue_k), data_ijk in self.facet_data():
# If this subset is null, move on
if not data_ijk.values.size:
continue
# Get the current axis
ax = self.facet_axis(row_i, col_j)
# Decide what color to plot with
kwargs["color"] = self._facet_color(hue_k, kw_color)
# Insert the other hue aesthetics if appropriate
for kw, val_list in self.hue_kws.items():
kwargs[kw] = val_list[hue_k]
# Insert a label in the keyword arguments for the legend
if self._hue_var is not None:
kwargs["label"] = utils.to_utf8(self.hue_names[hue_k])
# Get the actual data we are going to plot with
plot_data = data_ijk[list(args)]
if self._dropna:
plot_data = plot_data.dropna()
plot_args = [v for k, v in plot_data.iteritems()]
# Some matplotlib functions don't handle pandas objects correctly
if func_module.startswith("matplotlib"):
plot_args = [v.values for v in plot_args]
# Draw the plot
self._facet_plot(func, ax, plot_args, kwargs)
# Finalize the annotations and layout
self._finalize_grid(args[:2])
return self
def map_dataframe(self, func, *args, **kwargs):
"""Like ``.map`` but passes args as strings and inserts data in kwargs.
This method is suitable for plotting with functions that accept a
long-form DataFrame as a `data` keyword argument and access the
data in that DataFrame using string variable names.
Parameters
----------
func : callable
A plotting function that takes data and keyword arguments. Unlike
the `map` method, a function used here must "understand" Pandas
objects. It also must plot to the currently active matplotlib Axes
and take a `color` keyword argument. If faceting on the `hue`
dimension, it must also take a `label` keyword argument.
args : strings
Column names in self.data that identify variables with data to
plot. The data for each variable is passed to `func` in the
order the variables are specified in the call.
kwargs : keyword arguments
All keyword arguments are passed to the plotting function.
Returns
-------
self : object
Returns self.
"""
# If color was a keyword argument, grab it here
kw_color = kwargs.pop("color", None)
# Iterate over the data subsets
for (row_i, col_j, hue_k), data_ijk in self.facet_data():
# If this subset is null, move on
if not data_ijk.values.size:
continue
# Get the current axis
ax = self.facet_axis(row_i, col_j)
# Decide what color to plot with
kwargs["color"] = self._facet_color(hue_k, kw_color)
# Insert the other hue aesthetics if appropriate
for kw, val_list in self.hue_kws.items():
kwargs[kw] = val_list[hue_k]
# Insert a label in the keyword arguments for the legend
if self._hue_var is not None:
kwargs["label"] = self.hue_names[hue_k]
# Stick the facet dataframe into the kwargs
if self._dropna:
data_ijk = data_ijk.dropna()
kwargs["data"] = data_ijk
# Draw the plot
self._facet_plot(func, ax, args, kwargs)
# Finalize the annotations and layout
self._finalize_grid(args[:2])
return self
def _facet_color(self, hue_index, kw_color):
color = self._colors[hue_index]
if kw_color is not None:
return kw_color
elif color is not None:
return color
def _facet_plot(self, func, ax, plot_args, plot_kwargs):
# Draw the plot
func(*plot_args, **plot_kwargs)
# Sort out the supporting information
self._update_legend_data(ax)
self._clean_axis(ax)
def _finalize_grid(self, axlabels):
"""Finalize the annotations and layout."""
self.set_axis_labels(*axlabels)
self.set_titles()
self.fig.tight_layout()
def facet_axis(self, row_i, col_j):
"""Make the axis identified by these indices active and return it."""
# Calculate the actual indices of the axes to plot on
if self._col_wrap is not None:
ax = self.axes.flat[col_j]
else:
ax = self.axes[row_i, col_j]
# Get a reference to the axes object we want, and make it active
plt.sca(ax)
return ax
def despine(self, **kwargs):
"""Remove axis spines from the facets."""
utils.despine(self.fig, **kwargs)
return self
def set_axis_labels(self, x_var=None, y_var=None):
"""Set axis labels on the left column and bottom row of the grid."""
if x_var is not None:
self._x_var = x_var
self.set_xlabels(x_var)
if y_var is not None:
self._y_var = y_var
self.set_ylabels(y_var)
return self
def set_xlabels(self, label=None, **kwargs):
"""Label the x axis on the bottom row of the grid."""
if label is None:
label = self._x_var
for ax in self._bottom_axes:
ax.set_xlabel(label, **kwargs)
return self
def set_ylabels(self, label=None, **kwargs):
"""Label the y axis on the left column of the grid."""
if label is None:
label = self._y_var
for ax in self._left_axes:
ax.set_ylabel(label, **kwargs)
return self
def set_xticklabels(self, labels=None, step=None, **kwargs):
"""Set x axis tick labels of the grid."""
for ax in self.axes.flat:
if labels is None:
curr_labels = [l.get_text() for l in ax.get_xticklabels()]
if step is not None:
xticks = ax.get_xticks()[::step]
curr_labels = curr_labels[::step]
ax.set_xticks(xticks)
ax.set_xticklabels(curr_labels, **kwargs)
else:
ax.set_xticklabels(labels, **kwargs)
return self
def set_yticklabels(self, labels=None, **kwargs):
"""Set y axis tick labels on the left column of the grid."""
for ax in self.axes.flat:
if labels is None:
curr_labels = [l.get_text() for l in ax.get_yticklabels()]
ax.set_yticklabels(curr_labels, **kwargs)
else:
ax.set_yticklabels(labels, **kwargs)
return self
def set_titles(self, template=None, row_template=None, col_template=None,
**kwargs):
"""Draw titles either above each facet or on the grid margins.
Parameters
----------
template : string
Template for all titles with the formatting keys {col_var} and
{col_name} (if using a `col` faceting variable) and/or {row_var}
and {row_name} (if using a `row` faceting variable).
row_template:
Template for the row variable when titles are drawn on the grid
margins. Must have {row_var} and {row_name} formatting keys.
col_template:
Template for the row variable when titles are drawn on the grid
margins. Must have {col_var} and {col_name} formatting keys.
Returns
-------
self: object
Returns self.
"""
args = dict(row_var=self._row_var, col_var=self._col_var)
kwargs["size"] = kwargs.pop("size", mpl.rcParams["axes.labelsize"])
# Establish default templates
if row_template is None:
row_template = "{row_var} = {row_name}"
if col_template is None:
col_template = "{col_var} = {col_name}"
if template is None:
if self._row_var is None:
template = col_template
elif self._col_var is None:
template = row_template
else:
template = " | ".join([row_template, col_template])
row_template = utils.to_utf8(row_template)
col_template = utils.to_utf8(col_template)
template = utils.to_utf8(template)
if self._margin_titles:
if self.row_names is not None:
# Draw the row titles on the right edge of the grid
for i, row_name in enumerate(self.row_names):
ax = self.axes[i, -1]
args.update(dict(row_name=row_name))
title = row_template.format(**args)
bgcolor = self.fig.get_facecolor()
ax.annotate(title, xy=(1.02, .5), xycoords="axes fraction",
rotation=270, ha="left", va="center",
backgroundcolor=bgcolor, **kwargs)
if self.col_names is not None:
# Draw the column titles as normal titles
for j, col_name in enumerate(self.col_names):
args.update(dict(col_name=col_name))
title = col_template.format(**args)
self.axes[0, j].set_title(title, **kwargs)
return self
# Otherwise title each facet with all the necessary information
if (self._row_var is not None) and (self._col_var is not None):
for i, row_name in enumerate(self.row_names):
for j, col_name in enumerate(self.col_names):
args.update(dict(row_name=row_name, col_name=col_name))
title = template.format(**args)
self.axes[i, j].set_title(title, **kwargs)
elif self.row_names is not None and len(self.row_names):
for i, row_name in enumerate(self.row_names):
args.update(dict(row_name=row_name))
title = template.format(**args)