看完这篇文章，我彻底爱上了Python动态图表！

import pandas_alive
import pandas as pd

covid_df = pd.read_csv('data/covid19.csv', index_col=0, parse_dates=[0])
covid_df.plot_animated(filename='examples/example-barh-chart.gif', n_visible=15)

elec_df = pd.read_csv("data/Aus_Elec_Gen_1980_2018.csv", index_col=0, parse_dates=[0], thousands=',')
elec_df = elec_df.iloc[:20, :]
elec_df.fillna(0).plot_animated('examples/example-electricity-generated-australia.gif', period_fmt="%Y",
                                title='Australian Electricity Generation Sources 1980-2018')

covid_df = pd.read_csv('data/covid19.csv', index_col=0, parse_dates=[0])
covid_df.plot_animated(filename='examples/example-barv-chart.gif', orientation='v', n_visible=15)

covid_df = pd.read_csv('data/covid19.csv', index_col=0, parse_dates=[0])
covid_df.diff().fillna(0).plot_animated(filename='examples/example-line-chart.gif', kind='line', period_label={'x': 0.25, 'y': 0.9})

covid_df = pd.read_csv('data/covid19.csv', index_col=0, parse_dates=[0])
covid_df.sum(axis=1).fillna(0).plot_animated(filename='examples/example-bar-chart.gif', kind='bar',
        period_label={'x': 0.1, 'y': 0.9},
        enable_progress_bar=True, steps_per_period=2, interpolate_period=True, period_length=200
)

max_temp_df = pd.read_csv(
    "data/Newcastle_Australia_Max_Temps.csv",
    parse_dates={"Timestamp": ["Year", "Month", "Day"]},
)
min_temp_df = pd.read_csv(
    "data/Newcastle_Australia_Min_Temps.csv",
    parse_dates={"Timestamp": ["Year", "Month", "Day"]},
)

max_temp_df = max_temp_df.iloc[:5000, :]
min_temp_df = min_temp_df.iloc[:5000, :]

merged_temp_df = pd.merge_asof(max_temp_df, min_temp_df, on="Timestamp")
merged_temp_df.index = pd.to_datetime(merged_temp_df["Timestamp"].dt.strftime('%Y/%m/%d'))

keep_columns = ["Minimum temperature (Degree C)", "Maximum temperature (Degree C)"]
merged_temp_df[keep_columns].resample("Y").mean().plot_animated(filename='examples/example-scatter-chart.gif', kind="scatter",
                                                                title='Max & Min Temperature Newcastle, Australia')

covid_df = pd.read_csv('data/covid19.csv', index_col=0, parse_dates=[0])
covid_df.plot_animated(filename='examples/example-pie-chart.gif', kind="pie",
                       rotatelabels=True, period_label={'x': 0, 'y': 0})

multi_index_df = pd.read_csv("data/multi.csv", header=[0, 1], index_col=0)
multi_index_df.index = pd.to_datetime(multi_index_df.index, dayfirst=True)

map_chart = multi_index_df.plot_animated(
    kind="bubble",
    filename="examples/example-bubble-chart.gif",
    x_data_label="Longitude",
    y_data_label="Latitude",
    size_data_label="Cases",
    color_data_label="Cases",
    vmax=5, steps_per_period=3, interpolate_period=True, period_length=500,
    dpi=100
)

import geopandas
import pandas_alive
import contextily

gdf = geopandas.read_file('data/nsw-covid19-cases-by-postcode.gpkg')
gdf.index = gdf.postcode
gdf = gdf.drop('postcode',axis=1)

result = gdf.iloc[:, :20]
result['geometry'] = gdf.iloc[:, -1:]['geometry']

map_chart = result.plot_animated(filename='examples/example-geo-point-chart.gif',
                                 basemap_format={'source':contextily.providers.Stamen.Terrain})

import geopandas
import pandas_alive
import contextily

gdf = geopandas.read_file('data/italy-covid-region.gpkg')
gdf.index = gdf.region
gdf = gdf.drop('region',axis=1)

result = gdf.iloc[:, :20]
result['geometry'] = gdf.iloc[:, -1:]['geometry']

map_chart = result.plot_animated(filename='examples/example-geo-polygon-chart.gif',
                                 basemap_format={'source': contextily.providers.Stamen.Terrain})

covid_df = pd.read_csv('data/covid19.csv', index_col=0, parse_dates=[0])

animated_line_chart = covid_df.diff().fillna(0).plot_animated(kind='line', period_label=False,add_legend=False)
animated_bar_chart = covid_df.plot_animated(n_visible=10)

pandas_alive.animate_multiple_plots('examples/example-bar-and-line-chart.gif',
                                    [animated_bar_chart, animated_line_chart], enable_progress_bar=True)

def population():
    urban_df = pd.read_csv("data/urban_pop.csv", index_col=0, parse_dates=[0])

    animated_line_chart = (
        urban_df.sum(axis=1)
            .pct_change()
            .fillna(method='bfill')
            .mul(100)
            .plot_animated(kind="line", title="Total % Change in Population", period_label=False, add_legend=False)
    )

    animated_bar_chart = urban_df.plot_animated(n_visible=10, title='Top 10 Populous Countries', period_fmt="%Y")

    pandas_alive.animate_multiple_plots('examples/example-bar-and-line-urban-chart.gif',
                                        [animated_bar_chart, animated_line_chart],
                                        title='Urban Population 1977 - 2018', adjust_subplot_top=0.85,
                                        enable_progress_bar=True)

def life():
    data_raw = pd.read_csv("data/long.csv")

    list_G7 = [
        "Canada",
        "France",
        "Germany",
        "Italy",
        "Japan",
        "United Kingdom",
        "United States",
    ]

    data_raw = data_raw.pivot(
        index="Year", columns="Entity", values="Life expectancy (Gapminder, UN)"
    )

    data = pd.DataFrame()
    data["Year"] = data_raw.reset_index()["Year"]
    for country in list_G7:
        data[country] = data_raw[country].values

    data = data.fillna(method="pad")
    data = data.fillna(0)
    data = data.set_index("Year").loc[1900:].reset_index()

    data["Year"] = pd.to_datetime(data.reset_index()["Year"].astype(str))

    data = data.set_index("Year")
    data = data.iloc[:25, :]

    animated_bar_chart = data.plot_animated(
        period_fmt="%Y", perpendicular_bar_func="mean", period_length=200, fixed_max=True
    )

    animated_line_chart = data.plot_animated(
        kind="line", period_fmt="%Y", period_length=200, fixed_max=True
    )

    pandas_alive.animate_multiple_plots(
        "examples/life-expectancy.gif",
        plots=[animated_bar_chart, animated_line_chart],
        title="Life expectancy in G7 countries up to 2015",
        adjust_subplot_left=0.2, adjust_subplot_top=0.9, enable_progress_bar=True
    )

def nsw():
    import geopandas
    import pandas as pd
    import pandas_alive
    import contextily
    import matplotlib.pyplot as plt
    import json

    with open('data/package_show.json', 'r', encoding='utf8')as fp:
        data = json.load(fp)

    # Extract url to csv component
    covid_nsw_data_url = data["result"]["resources"][0]["url"]
    print(covid_nsw_data_url)

    # Read csv from data API url
    nsw_covid = pd.read_csv('data/confirmed_cases_table1_location.csv')
    postcode_dataset = pd.read_csv("data/postcode-data.csv")

    # Prepare data from NSW health dataset

    nsw_covid = nsw_covid.fillna(9999)
    nsw_covid["postcode"] = nsw_covid["postcode"].astype(int)

    grouped_df = nsw_covid.groupby(["notification_date", "postcode"]).size()
    grouped_df = pd.DataFrame(grouped_df).unstack()
    grouped_df.columns = grouped_df.columns.droplevel().astype(str)

    grouped_df = grouped_df.fillna(0)
    grouped_df.index = pd.to_datetime(grouped_df.index)

    cases_df = grouped_df

    # Clean data in postcode dataset prior to matching

    grouped_df = grouped_df.T
    postcode_dataset = postcode_dataset[postcode_dataset['Longitude'].notna()]
    postcode_dataset = postcode_dataset[postcode_dataset['Longitude'] != 0]
    postcode_dataset = postcode_dataset[postcode_dataset['Latitude'].notna()]
    postcode_dataset = postcode_dataset[postcode_dataset['Latitude'] != 0]
    postcode_dataset['Postcode'] = postcode_dataset['Postcode'].astype(str)

    # Build GeoDataFrame from Lat Long dataset and make map chart
    grouped_df['Longitude'] = grouped_df.index.map(postcode_dataset.set_index('Postcode')['Longitude'].to_dict())
    grouped_df['Latitude'] = grouped_df.index.map(postcode_dataset.set_index('Postcode')['Latitude'].to_dict())
    gdf = geopandas.GeoDataFrame(
        grouped_df, geometry=geopandas.points_from_xy(grouped_df.Longitude, grouped_df.Latitude), crs="EPSG:4326")
    gdf = gdf.dropna()

    # Prepare GeoDataFrame for writing to geopackage
    gdf = gdf.drop(['Longitude', 'Latitude'], axis=1)
    gdf.columns = gdf.columns.astype(str)
    gdf['postcode'] = gdf.index
    # gdf.to_file("data/nsw-covid19-cases-by-postcode.gpkg", layer='nsw-postcode-covid', driver="GPKG")

    # Prepare GeoDataFrame for plotting
    gdf.index = gdf.postcode
    gdf = gdf.drop('postcode', axis=1)
    gdf = gdf.to_crs("EPSG:3857")  # Web Mercator

    result = gdf.iloc[:, :22]
    result['geometry'] = gdf.iloc[:, -1:]['geometry']
    gdf = result

    map_chart = gdf.plot_animated(basemap_format={'source': contextily.providers.Stamen.Terrain}, cmap='cool')

    # cases_df.to_csv('data/nsw-covid-cases-by-postcode.csv')
    cases_df = cases_df.iloc[:22, :]

    from datetime import datetime

    bar_chart = cases_df.sum(axis=1).plot_animated(
        kind='line',
        label_events={
            'Ruby Princess Disembark': datetime.strptime("19/03/2020", "%d/%m/%Y"),
            # 'Lockdown': datetime.strptime("31/03/2020", "%d/%m/%Y")
        },
        fill_under_line_color="blue",
        add_legend=False
    )

    map_chart.ax.set_title('Cases by Location')

    grouped_df = pd.read_csv('data/nsw-covid-cases-by-postcode.csv', index_col=0, parse_dates=[0])
    grouped_df = grouped_df.iloc[:22, :]

    line_chart = (
        grouped_df.sum(axis=1)
            .cumsum()
            .fillna(0)
            .plot_animated(kind="line", period_label=False, title="Cumulative Total Cases", add_legend=False)
    )

    def current_total(values):
        total = values.sum()
        s = f'Total : {int(total)}'
        return {'x': .85, 'y': .2, 's': s, 'ha': 'right', 'size': 11}

    race_chart = grouped_df.cumsum().plot_animated(
        n_visible=5, title="Cases by Postcode", period_label=False, period_summary_func=current_total
    )

    import time

    timestr = time.strftime("%d/%m/%Y")

    plots = [bar_chart, line_chart, map_chart, race_chart]

    from matplotlib import rcParams

    rcParams.update({"figure.autolayout": False})
    # make sure figures are `Figure()` instances
    figs = plt.Figure()
    gs = figs.add_gridspec(2, 3, hspace=0.5)
    f3_ax1 = figs.add_subplot(gs[0, :])
    f3_ax1.set_title(bar_chart.title)
    bar_chart.ax = f3_ax1

    f3_ax2 = figs.add_subplot(gs[1, 0])
    f3_ax2.set_title(line_chart.title)
    line_chart.ax = f3_ax2

    f3_ax3 = figs.add_subplot(gs[1, 1])
    f3_ax3.set_title(map_chart.title)
    map_chart.ax = f3_ax3

    f3_ax4 = figs.add_subplot(gs[1, 2])
    f3_ax4.set_title(race_chart.title)
    race_chart.ax = f3_ax4

    timestr = cases_df.index.max().strftime("%d/%m/%Y")
    figs.suptitle(f"NSW COVID-19 Confirmed Cases up to {timestr}")

    pandas_alive.animate_multiple_plots(
        'examples/nsw-covid.gif',
        plots,
        figs,
        enable_progress_bar=True
    )

def italy():
    import geopandas
    import pandas as pd
    import pandas_alive
    import contextily
    import matplotlib.pyplot as plt

    region_gdf = geopandas.read_file('data/geo-data/italy-with-regions')
    region_gdf.NOME_REG = region_gdf.NOME_REG.str.lower().str.title()
    region_gdf = region_gdf.replace('Trentino-Alto Adige/Sudtirol', 'Trentino-Alto Adige')
    region_gdf = region_gdf.replace("Valle D'Aosta/Vallée D'Aoste\r\nValle D'Aosta/Vallée D'Aoste", "Valle d'Aosta")

    italy_df = pd.read_csv('data/Regional Data - Sheet1.csv', index_col=0, header=1, parse_dates=[0])

    italy_df = italy_df[italy_df['Region'] != 'NA']

    cases_df = italy_df.iloc[:, :3]
    cases_df['Date'] = cases_df.index
    pivoted = cases_df.pivot(values='New positives', index='Date', columns='Region')
    pivoted.columns = pivoted.columns.astype(str)
    pivoted = pivoted.rename(columns={'nan': 'Unknown Region'})

    cases_gdf = pivoted.T
    cases_gdf['geometry'] = cases_gdf.index.map(region_gdf.set_index('NOME_REG')['geometry'].to_dict())

    cases_gdf = cases_gdf[cases_gdf['geometry'].notna()]

    cases_gdf = geopandas.GeoDataFrame(cases_gdf, crs=region_gdf.crs, geometry=cases_gdf.geometry)

    gdf = cases_gdf

    result = gdf.iloc[:, :22]
    result['geometry'] = gdf.iloc[:, -1:]['geometry']
    gdf = result

    map_chart = gdf.plot_animated(basemap_format={'source': contextily.providers.Stamen.Terrain}, cmap='viridis')

    cases_df = pivoted
    cases_df = cases_df.iloc[:22, :]

    from datetime import datetime

    bar_chart = cases_df.sum(axis=1).plot_animated(
        kind='line',
        label_events={
            'Schools Close': datetime.strptime("4/03/2020", "%d/%m/%Y"),
            'Phase I Lockdown': datetime.strptime("11/03/2020", "%d/%m/%Y"),
            # '1M Global Cases': datetime.strptime("02/04/2020", "%d/%m/%Y"),
            # '100k Global Deaths': datetime.strptime("10/04/2020", "%d/%m/%Y"),
            # 'Manufacturing Reopens': datetime.strptime("26/04/2020", "%d/%m/%Y"),
            # 'Phase II Lockdown': datetime.strptime("4/05/2020", "%d/%m/%Y"),
        },
        fill_under_line_color="blue",
        add_legend=False
    )

    map_chart.ax.set_title('Cases by Location')

    line_chart = (
        cases_df.sum(axis=1)
            .cumsum()
            .fillna(0)
            .plot_animated(kind="line", period_label=False, title="Cumulative Total Cases", add_legend=False)
    )

    def current_total(values):
        total = values.sum()
        s = f'Total : {int(total)}'
        return {'x': .85, 'y': .1, 's': s, 'ha': 'right', 'size': 11}

    race_chart = cases_df.cumsum().plot_animated(
        n_visible=5, title="Cases by Region", period_label=False, period_summary_func=current_total
    )

    import time

    timestr = time.strftime("%d/%m/%Y")

    plots = [bar_chart, race_chart, map_chart, line_chart]

    # Otherwise titles overlap and adjust_subplot does nothing
    from matplotlib import rcParams
    from matplotlib.animation import FuncAnimation

    rcParams.update({"figure.autolayout": False})
    # make sure figures are `Figure()` instances
    figs = plt.Figure()
    gs = figs.add_gridspec(2, 3, hspace=0.5)
    f3_ax1 = figs.add_subplot(gs[0, :])
    f3_ax1.set_title(bar_chart.title)
    bar_chart.ax = f3_ax1

    f3_ax2 = figs.add_subplot(gs[1, 0])
    f3_ax2.set_title(race_chart.title)
    race_chart.ax = f3_ax2

    f3_ax3 = figs.add_subplot(gs[1, 1])
    f3_ax3.set_title(map_chart.title)
    map_chart.ax = f3_ax3

    f3_ax4 = figs.add_subplot(gs[1, 2])
    f3_ax4.set_title(line_chart.title)
    line_chart.ax = f3_ax4

    axes = [f3_ax1, f3_ax2, f3_ax3, f3_ax4]
    timestr = cases_df.index.max().strftime("%d/%m/%Y")
    figs.suptitle(f"Italy COVID-19 Confirmed Cases up to {timestr}")

    pandas_alive.animate_multiple_plots(
        'examples/italy-covid.gif',
        plots,
        figs,
        enable_progress_bar=True
    )

def simple():
    import pandas as pd
    import matplotlib.pyplot as plt
    import pandas_alive
    import numpy as np

    # Physical constants
    g = 9.81
    L = .4
    mu = 0.2

    THETA_0 = np.pi * 70 / 180  # init angle = 70degs
    THETA_DOT_0 = 0  # no init angVel
    DELTA_T = 0.01  # time stepping
    T = 1.5  # time period

    # Definition of ODE (ordinary differential equation)
    def get_theta_double_dot(theta, theta_dot):
        return -mu * theta_dot - (g / L) * np.sin(theta)

    # Solution to the differential equation
    def pendulum(t):
        # initialise changing values
        theta = THETA_0
        theta_dot = THETA_DOT_0
        delta_t = DELTA_T
        ang = []
        ang_vel = []
        ang_acc = []
        times = []
        for time in np.arange(0, t, delta_t):
            theta_double_dot = get_theta_double_dot(
                theta, theta_dot
            )
            theta += theta_dot * delta_t
            theta_dot += theta_double_dot * delta_t
            times.append(time)
            ang.append(theta)
            ang_vel.append(theta_dot)
            ang_acc.append(theta_double_dot)
        data = np.array([ang, ang_vel, ang_acc])
        return pd.DataFrame(data=data.T, index=np.array(times), columns=["angle", "ang_vel", "ang_acc"])

    # units used for ref: ["angle [rad]", "ang_vel [rad/s]", "ang_acc [rad/s^2]"]
    df = pendulum(T)
    df.index.names = ["Time (s)"]
    print(df)

    # generate dataFrame for animated bubble plot
    df2 = pd.DataFrame(index=df.index)
    df2["dx (m)"] = L * np.sin(df["angle"])
    df2["dy (m)"] = -L * np.cos(df["angle"])
    df2["ang_vel"] = abs(df["ang_vel"])
    df2["size"] = df2["ang_vel"] * 100  # scale angular vels to get nice size on bubble plot
    print(df2)

    # static pandas plots
    #
    # print(plt.style.available)
    # NOTE: 2 lines below required in Jupyter to switch styles correctly
    plt.rcParams.update(plt.rcParamsDefault)
    plt.style.use("ggplot")  # set plot style

    fig, (ax1a, ax2b) = plt.subplots(1, 2, figsize=(8, 4), dpi=100)  # 1 row, 2 subplots
    # fig.subplots_adjust(wspace=0.1)      # space subplots in row
    fig.set_tight_layout(True)
    fontsize = "small"

    df.plot(ax=ax1a).legend(fontsize=fontsize)
    ax1a.set_title("Outputs vs Time", fontsize="medium")
    ax1a.set_xlabel('Time [s]', fontsize=fontsize)
    ax1a.set_ylabel('Amplitudes', fontsize=fontsize);

    df.plot(ax=ax2b, x="angle", y=["ang_vel", "ang_acc"]).legend(fontsize=fontsize)
    ax2b.set_title("Outputs vs Angle | Phase-Space", fontsize="medium")
    ax2b.set_xlabel('Angle [rad]', fontsize=fontsize)
    ax2b.set_ylabel('Angular Velocity / Acc', fontsize=fontsize)

    # sample scatter plot with colorbar
    fig, ax = plt.subplots()
    sc = ax.scatter(df2["dx (m)"], df2["dy (m)"], s=df2["size"] * .1, c=df2["ang_vel"], cmap="jet")
    cbar = fig.colorbar(sc)
    cbar.set_label(label="ang_vel [rad/s]", fontsize="small")
    # sc.set_clim(350, 400)
    ax.tick_params(labelrotation=0, labelsize="medium")
    ax_scale = 1.
    ax.set_xlim(-L * ax_scale, L * ax_scale)
    ax.set_ylim(-L * ax_scale - 0.1, L * ax_scale - 0.1)
    # make axes square: a circle shows as a circle
    ax.set_aspect(1 / ax.get_data_ratio())
    ax.arrow(0, 0, df2["dx (m)"].iloc[-1], df2["dy (m)"].iloc[-1],
             color="dimgray", ls=":", lw=2.5, width=.0, head_width=0, zorder=-1
             )
    ax.text(0, 0.15, s="size and colour of pendulum bob\nbased on pd column\nfor angular velocity",
            ha='center', va='center')

    # plt.show()

    dpi = 100
    ax_scale = 1.1
    figsize = (3, 3)
    fontsize = "small"

    # set up figure to pass onto `pandas_alive`
    # NOTE: by using Figure (capital F) instead of figure() `FuncAnimation` seems to run twice as fast!
    # fig1, ax1 = plt.subplots()
    fig1 = plt.Figure()
    ax1 = fig1.add_subplot()
    fig1.set_size_inches(figsize)
    ax1.set_title("Simple pendulum animation, L=" + str(L) + "m", fontsize="medium")
    ax1.set_xlabel("Time (s)", color='dimgray', fontsize=fontsize)
    ax1.set_ylabel("Amplitudes", color='dimgray', fontsize=fontsize)
    ax1.tick_params(labelsize=fontsize)

    # pandas_alive
    line_chart = df.plot_animated(filename="pend-line.gif", kind='line', period_label={'x': 0.05, 'y': 0.9},
                                  steps_per_period=1, interpolate_period=False, period_length=50,
                                  period_fmt='Time:{x:10.2f}',
                                  enable_progress_bar=True, fixed_max=True, dpi=100, fig=fig1
                                  )
    plt.close()

    # Video('examples/pend-line.mp4', html_attributes="controls muted autoplay")

    # set up and generate animated scatter plot
    #

    # set up figure to pass onto `pandas_alive`
    # NOTE: by using Figure (capital F) instead of figure() `FuncAnimation` seems to run twice as fast!
    fig1sc = plt.Figure()
    ax1sc = fig1sc.add_subplot()
    fig1sc.set_size_inches(figsize)
    ax1sc.set_title("Simple pendulum animation, L=" + str(L) + "m", fontsize="medium")
    ax1sc.set_xlabel("Time (s)", color='dimgray', fontsize=fontsize)
    ax1sc.set_ylabel("Amplitudes", color='dimgray', fontsize=fontsize)
    ax1sc.tick_params(labelsize=fontsize)

    # pandas_alive
    scatter_chart = df.plot_animated(filename="pend-scatter.gif", kind='scatter', period_label={'x': 0.05, 'y': 0.9},
                                     steps_per_period=1, interpolate_period=False, period_length=50,
                                     period_fmt='Time:{x:10.2f}',
                                     enable_progress_bar=True, fixed_max=True, dpi=100, fig=fig1sc, size="ang_vel"
                                     )
    plt.close()

    print("Points size follows one of the pd columns: ang_vel")
    # Video('./pend-scatter.gif', html_attributes="controls muted autoplay")

    # set up and generate animated bar race chart
    #
    # set up figure to pass onto `pandas_alive`
    # NOTE: by using Figure (capital F) instead of figure() `FuncAnimation` seems to run twice as fast!
    fig2 = plt.Figure()
    ax2 = fig2.add_subplot()
    fig2.set_size_inches(figsize)
    ax2.set_title("Simple pendulum animation, L=" + str(L) + "m", fontsize="medium")
    ax2.set_xlabel("Amplitudes", color='dimgray', fontsize=fontsize)
    ax2.set_ylabel("", color='dimgray', fontsize="x-small")
    ax2.tick_params(labelsize=fontsize)

    # pandas_alive
    race_chart = df.plot_animated(filename="pend-race.gif", kind='race', period_label={'x': 0.05, 'y': 0.9},
                                  steps_per_period=1, interpolate_period=False, period_length=50,
                                  period_fmt='Time:{x:10.2f}',
                                  enable_progress_bar=True, fixed_max=False, dpi=100, fig=fig2
                                  )
    plt.close()

    # set up and generate bubble animated plot
    #

    # set up figure to pass onto `pandas_alive`
    # NOTE: by using Figure (capital F) instead of figure() `FuncAnimation` seems to run twice as fast!
    fig3 = plt.Figure()
    ax3 = fig3.add_subplot()
    fig3.set_size_inches(figsize)
    ax3.set_title("Simple pendulum animation, L=" + str(L) + "m", fontsize="medium")
    ax3.set_xlabel("Hor Displacement (m)", color='dimgray', fontsize=fontsize)
    ax3.set_ylabel("Ver Displacement (m)", color='dimgray', fontsize=fontsize)
    # limits & ratio below get the graph square
    ax3.set_xlim(-L * ax_scale, L * ax_scale)
    ax3.set_ylim(-L * ax_scale - 0.1, L * ax_scale - 0.1)
    ratio = 1.  # this is visual ratio of axes
    ax3.set_aspect(ratio / ax3.get_data_ratio())

    ax3.arrow(0, 0, df2["dx (m)"].iloc[-1], df2["dy (m)"].iloc[-1],
              color="dimgray", ls=":", lw=1, width=.0, head_width=0, zorder=-1)

    # pandas_alive
    bubble_chart = df2.plot_animated(
        kind="bubble", filename="pend-bubble.gif",
        x_data_label="dx (m)", y_data_label="dy (m)",
        size_data_label="size", color_data_label="ang_vel", cmap="jet",
        period_label={'x': 0.05, 'y': 0.9}, vmin=None, vmax=None,
        steps_per_period=1, interpolate_period=False, period_length=50, period_fmt='Time:{x:10.2f}s',
        enable_progress_bar=True, fixed_max=False, dpi=dpi, fig=fig3
    )
    plt.close()

    print("Bubble size & colour animates with pd data column for ang_vel.")

    # Combined plots
    #
    fontsize = "x-small"
    # Otherwise titles overlap and subplots_adjust does nothing
    from matplotlib import rcParams
    rcParams.update({"figure.autolayout": False})

    figs = plt.Figure(figsize=(9, 4), dpi=100)
    figs.subplots_adjust(wspace=0.1)
    gs = figs.add_gridspec(2, 2)

    ax1 = figs.add_subplot(gs[0, 0])
    ax1.set_xlabel("Time(s)", color='dimgray', fontsize=fontsize)
    ax1.set_ylabel("Amplitudes", color='dimgray', fontsize=fontsize)
    ax1.tick_params(labelsize=fontsize)

    ax2 = figs.add_subplot(gs[1, 0])
    ax2.set_xlabel("Amplitudes", color='dimgray', fontsize=fontsize)
    ax2.set_ylabel("", color='dimgray', fontsize=fontsize)
    ax2.tick_params(labelsize=fontsize)

    ax3 = figs.add_subplot(gs[:, 1])
    ax3.set_xlabel("Hor Displacement (m)", color='dimgray', fontsize=fontsize)
    ax3.set_ylabel("Ver Displacement (m)", color='dimgray', fontsize=fontsize)
    ax3.tick_params(labelsize=fontsize)
    # limits & ratio below get the graph square
    ax3.set_xlim(-L * ax_scale, L * ax_scale)
    ax3.set_ylim(-L * ax_scale - 0.1, L * ax_scale - 0.1)
    ratio = 1.  # this is visual ratio of axes
    ax3.set_aspect(ratio / ax3.get_data_ratio())

    line_chart.ax = ax1
    race_chart.ax = ax2
    bubble_chart.ax = ax3

    plots = [line_chart, race_chart, bubble_chart]
    # pandas_alive combined using custom figure
    pandas_alive.animate_multiple_plots(
        filename='pend-combined.gif', plots=plots, custom_fig=figs, dpi=100, enable_progress_bar=True,
        adjust_subplot_left=0.2, adjust_subplot_right=None,
        title="Simple pendulum animations, L=" + str(L) + "m", title_fontsize="medium"
    )
    plt.close()

# 中文显示
plt.rcParams['font.sans-serif'] = ['SimHei']  # Windows
plt.rcParams['font.sans-serif'] = ['Hiragino Sans GB'] # Mac
plt.rcParams['axes.unicode_minus'] = False

# 读取数据
df_result = pd.read_csv('data/yuhuanshui.csv', index_col=0, parse_dates=[0])
# 生成图表
animated_line_chart = df_result.diff().fillna(0).plot_animated(kind='line', period_label=False, add_legend=False)
animated_bar_chart = df_result.plot_animated(n_visible=10)
pandas_alive.animate_multiple_plots('examples/yuhuanshui.gif',
                                    [animated_bar_chart, animated_line_chart], enable_progress_bar=True,
                                    title='我是余欢水演职人员热度排行')