import matplotlib.pyplot as plt
from pandas.api.types import is_datetime64_any_dtype
from trackintel.visualization.util import regular_figure, save_fig
[docs]def plot_modal_split(
df_modal_split_in,
out_path=None,
date_fmt_x_axis="%W",
fig=None,
axis=None,
title=None,
x_label=None,
y_label=None,
x_pad=10,
y_pad=10,
title_pad=1.02,
skip_xticks=0,
n_col_legend=5,
borderaxespad=0.5,
):
"""
Plot modal split as returned by `trackintel.analysis.modal_split.calculate_modal_split`
Parameters
----------
df_modal_split : DataFrame
DataFrame with modal split information. Format is
out_path : str, optional
Path to store the figure
date_fmt_x_axis : str, default: '%W'
strftime() date format code that is used for the x-axis
title : str, optional
x_label : str, optional
y_label : str, optional
fig : matplotlib.figure
Only used if axis is provided as well.
axis : matplotlib axes
x_pad : float, default: 10
Used to set ax.xaxis.labelpad
y_pad : float, default: 10
Used to set ax.yaxis.labelpad
title_pad : float, default: 1.02
Passed on to `matplotlib.pyplot.title`
skip_xticks : int, default: 1
Every nth x-tick label is kept.
n_col_legend : int
Passed on as `ncol` to matplotlib.pyplot.legend()
borderaxespad : float
Passed on to matplotlib.pyplot.legend()
Returns
-------
fig : Matplotlib figure handle
ax : Matplotlib axis handle
Examples
--------
>>> modal_split = calculate_modal_split(triplegs, metric='count', freq='D', per_user=False)
>>> plot_modal_split(modal_split, out_path=tmp_file, date_fmt_x_axis='%d',
>>> y_label='Percentage of daily count', x_label='days')
"""
df_modal_split = df_modal_split_in.copy()
if axis is None:
fig, ax = regular_figure()
else:
ax = axis
# make sure that modal split is only of a single user
if isinstance(df_modal_split.index[0], tuple):
raise ValueError(
"This function can not support multiindex types. Use 'pandas.MultiIndex.droplevel' or pass "
"the `per_user=False` flag in 'calculate_modal_split' function."
)
if not is_datetime64_any_dtype(df_modal_split.index.dtype):
raise ValueError(
"Index of modal split has to be a datetime type. This problem can be solved if the 'freq' "
"keyword of 'calculate_modal_split is not None'"
)
# set date formatter
df_modal_split.index = df_modal_split.index.map(lambda s: s.strftime(date_fmt_x_axis))
# plotting
df_modal_split.plot.bar(stacked=True, ax=ax)
# skip ticks for X axis
if skip_xticks > 0:
for i, tick in enumerate(ax.xaxis.get_major_ticks()):
if i % skip_xticks != 0:
tick.set_visible(False)
# We use a nice trick to put the legend out of the plot and to scale it automatically
# https://stackoverflow.com/questions/4700614/how-to-put-the-legend-out-of-the-plot
box = ax.get_position()
ax.set_position([box.x0, box.y0 + box.height * 0.1, box.width, box.height * 0.9])
# Put a legend below current axis
ax.legend(
loc="upper center",
bbox_to_anchor=(0.5, -0.05),
fancybox=True,
frameon=False,
ncol=n_col_legend,
borderaxespad=borderaxespad,
)
if title is not None:
ax.set_title(title, y=title_pad)
ax.set_xlabel(x_label)
ax.set_ylabel(y_label)
if fig is not None:
fig.autofmt_xdate()
plt.tight_layout()
ax.xaxis.labelpad = x_pad
ax.yaxis.labelpad = y_pad
if out_path is not None:
save_fig(out_path)
return fig, ax