Plot filled area charts to visualise distribution in time series data

LICENSE

MIT License

Copyright (c) 2019 Stefan Pfenninger

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

plot.py

import matplotlib.pyplot as plt
import matplotlib.colors as mcolors
import matplotlib.patches as mpatches
import seaborn as sns


def draw_areas(df, legend=True, ax=None, pal=None, lw=2.5, alpha=1.0, only_one=False, **kwargs):
    if not pal:
        pal = sns.color_palette("Reds_r")

    data = df.describe().T

    data['90%'] = df.quantile(0.9)
    data['10%'] = df.quantile(0.1)

    data = data.rename(columns={'50%': 'Median'})

    ax = data.loc[:, 'Median'].plot(
        lw=lw,
        color=mcolors.colorConverter.to_rgb(pal[0]),
        alpha=alpha, ax=ax)

    if only_one:
        ax.fill_between(range(len(data)), data['25%'], data['75%'], color=pal[2], alpha=alpha)
    else:
        ax.fill_between(range(len(data)), data['min'], data['max'], color=pal[3], alpha=alpha)
        ax.fill_between(range(len(data)), data['10%'], data['90%'], color=pal[2], alpha=alpha)
        ax.fill_between(range(len(data)), data['25%'], data['75%'], color=pal[1], alpha=alpha)

    # Add legend entries
    if only_one:
        ax.add_patch(plt.Rectangle((0, 0), 0, 0, facecolor=pal[2], alpha=alpha, label='25% - 75%'))
    else:
        ax.add_patch(plt.Rectangle((0, 0), 0, 0, facecolor=pal[0], alpha=alpha, label='25% - 75%'))
        ax.add_patch(plt.Rectangle((0, 0), 0, 0, facecolor=pal[1], alpha=alpha, label='10% - 90%'))
        ax.add_patch(plt.Rectangle((0, 0), 0, 0, facecolor=pal[2], alpha=alpha, label='Min - Max'))

    if legend:
        leg = ax.legend(
            loc='lower center',
            bbox_to_anchor=(0, 0.85, 1, 1),
            ncol=4)
        leg.draw_frame(False)

    ax.xaxis.get_major_ticks()[0].label1.set_visible(False)

    return ax


def downsample_df_with_mean(df, freq):
    s = (df.index.to_series() / freq).astype(int)
    return df.groupby(s).mean()


def plot(series, palette=None, ax=None, freq='1D', resample=True, only_one=False, full_monthly_labels=True, **kwargs):
    """`series` is a pandas.Series with hourly data."""
    df = series.to_frame()
    if resample:
        df = df.resample(freq).mean()

    # Get rid of Feb 29
    df = df[~((df.index.month == 2) & (df.index.day == 29))]

    df.columns = ['data']
    df['Year'] = df.index.year
    df['HourOfYear'] = ((df.index.dayofyear - 1) * 24) + df.index.hour

    df_plot = df.pivot(values='data', columns='Year', index='HourOfYear')
    df_plot.index.name = None

    freqs = {
        '1D': 24,
        '7D': 168,
        '1W': 168,
        '1M': 720,
    }

    df_plot = downsample_df_with_mean(df_plot, freqs[freq])

    # Reindex to integer
    df_plot.index = list(range(len(df_plot)))

    ax = draw_areas(df_plot.T, pal=palette, ax=ax, only_one=only_one, **kwargs)
    
    months = ['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']

    if freq == '1D':
        xticks = [0, 31, 59, 90, 120, 151, 181, 212, 242, 273, 303, 334]
        xlabels = months
        xticks_minor = xlabels_minor = None
    elif freq == '1W' or freq == '7D':
        xticks = [i / 7 for i in [0, 31, 59, 90, 120, 151, 181, 212, 242, 273, 303, 334]]
        xlabels = months
        xticks_minor = xlabels_minor = None
    elif freq == '1M':
        if full_monthly_labels:
            xticks = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
            xlabels = months
            xticks_minor = xlabels_minor = None
        else:
            xticks = [0, 3, 6, 9]
            xlabels = ['Jan', 'Apr', 'Jul', 'Oct']
            xticks_minor = [1, 2, 4, 5, 7, 8, 10, 11]
            xlabels_minor = [''] * len(xticks_minor)

    ax.set_xticks(xticks)
    ax.set_xticklabels(xlabels)
    if xticks_minor:
        ax.set_xticks(xticks_minor, minor=True)
        ax.set_xticklabels(xlabels_minor, minor=True)

    for label in ax.get_xticklabels():
        label.set_visible(True)

    ax.set_xlabel(None)

    return ax