ENH: Histogram axis labels

[sorenwacker]

Feature Type

• Adding new functionality to pandas

• Changing existing functionality in pandas

• Removing existing functionality in pandas

Problem Description

How about improving the labels on the generated histograms?

df = pd.DataFrame(dict(a=np.random.normal(size=100), b=np.random.normal(size=100)+100, c=np.random.choice(['A', 'B'], size=100)))
df

Using the series object, the figure has no labels at all.

df[['a']].hist()

When using the a dataframe objet, the subplots titles are the variable names.

Actually, the x-axis correspond to the values of variable a, instead of the title the name of the axis shoud be the x-axis label.
Since we have a histogram, the y-axes should be labeled with 'Counts'.

And using by= should add a legend to know which color is for what label.

df.hist(by='c')

This plot also should look more like this:

The labels do not make sense anymore since value from A and B

Feature Description

Instead of using title(column-name) use xlabel(column-name). Set 'Count' as y-label by default.

Alternative Solutions

Allow seaborn as backend?

Additional Context

No response

[topper-123]

Yes, the current output for frames could be improved...

[Rylie-W]

take

[Rylie-W]

Hello @topper-123 ,

I have two questions during the implementation:

1. According to the documentation of matplotlib.pyplot.hist(), setting the label parameter does not conflict with adding a legend. However, in pandas, it is not allowed to set both simultaneously. Some users may want to change the label without modifying the original DataFrame instance, so it might be necessary to implement this functionality.
   

   pandas/pandas/plotting/_matplotlib/hist.py

   Lines 482 to 483 in ef58af3

   	if legend and "label" in kwds:
   	raise ValueError("Cannot use both legend and label")

2. From the following two code snippets, pandas.DataFrame.hist() currently only supports drawing histograms for numeric or datetime data.
   

   pandas/pandas/plotting/_matplotlib/hist.py

   Lines 510 to 512 in ef58af3

   	data = data.select_dtypes(
   	include=(np.number, "datetime64", "datetimetz"), exclude="timedelta"
   	)

   
   

   pandas/pandas/plotting/_matplotlib/hist.py

   Lines 271 to 305 in ef58af3

   	numeric_only: bool = True,
   	figsize: tuple[float, float] | None = None,
   	sharex: bool = True,
   	sharey: bool = True,
   	layout=None,
   	rot: float = 0,
   	ax=None,
   	**kwargs,
   	):
   	# error: Non-overlapping equality check (left operand type: "Optional[Tuple[float,
   	# float]]", right operand type: "Literal['default']")
   	if figsize == "default": # type: ignore[comparison-overlap]
   	# allowed to specify mpl default with 'default'
   	raise ValueError(
   	"figsize='default' is no longer supported. "
   	"Specify figure size by tuple instead"
   	)
   	grouped = data.groupby(by)
   	if column is not None:
   	grouped = grouped[column]
   	naxes = len(grouped)
   	fig, axes = create_subplots(
   	naxes=naxes, figsize=figsize, sharex=sharex, sharey=sharey, ax=ax, layout=layout
   	)
   	_axes = flatten_axes(axes)
   	for i, (key, group) in enumerate(grouped):
   	ax = _axes[i]
   	if numeric_only and isinstance(group, ABCDataFrame):
   	group = group._get_numeric_data()
   	plotf(group, ax, **kwargs)
   	ax.set_title(pprint_thing(key))

   
   Should we support drawing histograms for non-numeric data, such as the column 'd', similar to matplotlib.pyplot.hist()? Otherwise, we should consider removing the numeric_only parameter, as it has a default value of True and cannot be specified by the user.

import pandas as pd
import numpy as np
import matplotlib as plt

df = pd.DataFrame(dict(a=np.random.normal(size=100), b=np.random.normal(size=100)+100, c=np.random.choice(['A', 'B'], size=100), d=np.random.choice(['A', 'B'], size=100)))
df.hist(by='c')
plt.pyplot.show()

plt.pyplot.hist(df.iloc[:,3])
plt.pyplot.show()