pandasのdataframe.plot.barとdataframe.plot(kind=bar)の違い

初心者です。
python3で、pandasを使ってデータフレームの可視化(業務書類に使用したいので、最終的にはかなり調整をして完成されたレイアウトにしていきたい)の勉強中です。

調べてみると、例えば棒グラフを書くときに、
df.plot.bar(stacked=1)
のようにも、
df.plot(kind=bar, stacked=1)
のようにも書けるようです。

やってみると見た目はまったく同じものができましたが、これはどういう違いがあるのでしょうか。
例えば、引数の一番最初のものに特権があってカッコの外に出せるという暗黙のルールがあるとか、別のメソッドで実は似ているけど違うコードで作られたグラフで用途により使いわけたほうがよいとか、そういうことをお尋ねしたいです。

宜しくお願いします。

行動規範の内容に同意します

回答1件

ベストアンサー

「引数の一番最初のものに特権があってカッコの外に出せるという暗黙のルール」はありません。
使ったことがないので詳しくはわかりませんが、df.plotでキーワード引数kind=barが指定されたときにはdf.plot.barを呼び出すように実装されている可能性が高いです。
kindに他のもの(histとかboxとか)を指定したときも同様です。

調べてみたところ、大本はpd.plotting._core.PlotAccessor.barのようです。df.plotでキーワード引数kind=barが指定されたときもdf.plot.barも、pd.plotting._core.PlotAccessor.barに帰着するので同じ動きをするということです。なぜそうなるかというと、pandasの作成者がそう作ったからであるとしか言いようがなく、一般的にどうこういえるものではありません。

それぞれの説明を見たければhelp(df.plot)とか、print(df.plot.doc)とかをやってみてください。
以下の様な説明が出ます。(英語を読むのが大変かもしれませんが)

python
1>>> print(df.plot.__doc__)
2
3    Make plots of Series or DataFrame.
4
5    Uses the backend specified by the
6    option ``plotting.backend``. By default, matplotlib is used.
7
8    Parameters
9    ----------
10    data : Series or DataFrame
11        The object for which the method is called.
12    x : label or position, default None
13        Only used if data is a DataFrame.
14    y : label, position or list of label, positions, default None
15        Allows plotting of one column versus another. Only used if data is a
16        DataFrame.
17    kind : str
18        The kind of plot to produce:
19
20        - 'line' : line plot (default)
21        - 'bar' : vertical bar plot
22        - 'barh' : horizontal bar plot
23        - 'hist' : histogram
24        - 'box' : boxplot
25        - 'kde' : Kernel Density Estimation plot
26        - 'density' : same as 'kde'
27        - 'area' : area plot
28        - 'pie' : pie plot
29        - 'scatter' : scatter plot
30        - 'hexbin' : hexbin plot.
31
32    figsize : a tuple (width, height) in inches
33    use_index : bool, default True
34        Use index as ticks for x axis.
35    title : str or list
36        Title to use for the plot. If a string is passed, print the string
37        at the top of the figure. If a list is passed and `subplots` is
38        True, print each item in the list above the corresponding subplot.
39    grid : bool, default None (matlab style default)
40        Axis grid lines.
41    legend : bool or {'reverse'}
42        Place legend on axis subplots.
43    style : list or dict
44        The matplotlib line style per column.
45    logx : bool or 'sym', default False
46        Use log scaling or symlog scaling on x axis.
47        .. versionchanged:: 0.25.0
48
49    logy : bool or 'sym' default False
50        Use log scaling or symlog scaling on y axis.
51        .. versionchanged:: 0.25.0
52
53    loglog : bool or 'sym', default False
54        Use log scaling or symlog scaling on both x and y axes.
55        .. versionchanged:: 0.25.0
56
57    xticks : sequence
58        Values to use for the xticks.
59    yticks : sequence
60        Values to use for the yticks.
61    xlim : 2-tuple/list
62    ylim : 2-tuple/list
63    rot : int, default None
64        Rotation for ticks (xticks for vertical, yticks for horizontal
65        plots).
66    fontsize : int, default None
67        Font size for xticks and yticks.
68    colormap : str or matplotlib colormap object, default None
69        Colormap to select colors from. If string, load colormap with that
70        name from matplotlib.
71    colorbar : bool, optional
72        If True, plot colorbar (only relevant for 'scatter' and 'hexbin'
73        plots).
74    position : float
75        Specify relative alignments for bar plot layout.
76        From 0 (left/bottom-end) to 1 (right/top-end). Default is 0.5
77        (center).
78    table : bool, Series or DataFrame, default False
79        If True, draw a table using the data in the DataFrame and the data
80        will be transposed to meet matplotlib's default layout.
81        If a Series or DataFrame is passed, use passed data to draw a
82        table.
83    yerr : DataFrame, Series, array-like, dict and str
84        See :ref:`Plotting with Error Bars <visualization.errorbars>` for
85        detail.
86    xerr : DataFrame, Series, array-like, dict and str
87        Equivalent to yerr.
88    mark_right : bool, default True
89        When using a secondary_y axis, automatically mark the column
90        labels with "(right)" in the legend.
91    include_bool : bool, default is False
92        If True, boolean values can be plotted.
93    backend : str, default None
94        Backend to use instead of the backend specified in the option
95        ``plotting.backend``. For instance, 'matplotlib'. Alternatively, to
96        specify the ``plotting.backend`` for the whole session, set
97        ``pd.options.plotting.backend``.
98
99        .. versionadded:: 1.0.0
100
101    **kwargs
102        Options to pass to matplotlib plotting method.
103
104    Returns
105    -------
106    :class:`matplotlib.axes.Axes` or numpy.ndarray of them
107        If the backend is not the default matplotlib one, the return value
108        will be the object returned by the backend.
109
110    Notes
111    -----
112    - See matplotlib documentation online for more on this subject
113    - If `kind` = 'bar' or 'barh', you can specify relative alignments
114      for bar plot layout by `position` keyword.
115      From 0 (left/bottom-end) to 1 (right/top-end). Default is 0.5
116      (center)

python
1>>> print(df.plot.bar.__doc__)
2
3        Vertical bar plot.
4
5        A bar plot is a plot that presents categorical data with
6        rectangular bars with lengths proportional to the values that they
7        represent. A bar plot shows comparisons among discrete categories. One
8        axis of the plot shows the specific categories being compared, and the
9        other axis represents a measured value.
10
11        Parameters
12        ----------
13        x : label or position, optional
14            Allows plotting of one column versus another. If not specified,
15            the index of the DataFrame is used.
16        y : label or position, optional
17            Allows plotting of one column versus another. If not specified,
18            all numerical columns are used.
19        **kwargs
20            Additional keyword arguments are documented in
21            :meth:`DataFrame.plot`.
22
23        Returns
24        -------
25        matplotlib.axes.Axes or np.ndarray of them
26            An ndarray is returned with one :class:`matplotlib.axes.Axes`
27            per column when ``subplots=True``.
28
29        See Also
30        --------
31        DataFrame.plot.barh : Horizontal bar plot.
32        DataFrame.plot : Make plots of a DataFrame.
33        matplotlib.pyplot.bar : Make a bar plot with matplotlib.
34
35        Examples
36        --------
37        Basic plot.
38
39        .. plot::
40            :context: close-figs
41
42            >>> df = pd.DataFrame({'lab':['A', 'B', 'C'], 'val':[10, 30, 20]})
43            >>> ax = df.plot.bar(x='lab', y='val', rot=0)
44
45        Plot a whole dataframe to a bar plot. Each column is assigned a
46        distinct color, and each row is nested in a group along the
47        horizontal axis.
48
49        .. plot::
50            :context: close-figs
51
52            >>> speed = [0.1, 17.5, 40, 48, 52, 69, 88]
53            >>> lifespan = [2, 8, 70, 1.5, 25, 12, 28]
54            >>> index = ['snail', 'pig', 'elephant',
55            ...          'rabbit', 'giraffe', 'coyote', 'horse']
56            >>> df = pd.DataFrame({'speed': speed,
57            ...                    'lifespan': lifespan}, index=index)
58            >>> ax = df.plot.bar(rot=0)
59
60        Instead of nesting, the figure can be split by column with
61        ``subplots=True``. In this case, a :class:`numpy.ndarray` of
62        :class:`matplotlib.axes.Axes` are returned.
63
64        .. plot::
65            :context: close-figs
66
67            >>> axes = df.plot.bar(rot=0, subplots=True)
68            >>> axes[1].legend(loc=2)  # doctest: +SKIP
69
70        Plot a single column.
71
72        .. plot::
73            :context: close-figs
74
75            >>> ax = df.plot.bar(y='speed', rot=0)
76
77        Plot only selected categories for the DataFrame.
78
79        .. plot::
80            :context: close-figs
81
82            >>> ax = df.plot.bar(x='lifespan', rot=0)