def plot_batch_results(df,
health_metrics=('Susceptible', 'Infected',
'Hospitalization', 'Severe',
'Recovered_Immune', 'Death'),
ecom_metrics=('Q1', 'Q2', 'Q3', 'Q4', 'Q5')):
"""
Plot the results of a batch executions contained in the given DataFrame
:param ecom_metrics:
:param health_metrics:
:param df: Pandas DataFrame returned by batch_experiment method
"""
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=[20, 5])
ax[0].set_title('Average Contagion Evolution')
ax[0].set_xlabel("Nº of Days")
ax[0].set_ylabel("% of Population")
for col in health_metrics:
means = df[(df["Metric"] == col)]['Avg'].values
std = df[(df["Metric"] == col)]['Std'].values
plot_mean_std(ax[0], means, std, legend=col, color=color1(col))
handles, labels = ax[0].get_legend_handles_labels()
lgd = ax[0].legend(handles, labels, loc='top right')
mmax = 0.0
mmin = np.inf
smax = 0
smin = np.inf
for col in ecom_metrics:
val = df[(df["Metric"] == col)]['Avg'].values
tmp = int(np.max(val))
mmax = np.max([mmax, tmp])
tmp = np.min(val)
mmin = np.min([mmin, tmp])
val = df[(df["Metric"] == col)]['Std'].values
tmp = np.max(val)
smax = np.max([smax, tmp])
tmp = np.min(val)
smin = np.min([smin, tmp])
ax[1].set_title('Average Economical Impact')
ax[1].set_xlabel("Nº of Days")
ax[1].set_ylabel("Wealth")
for col in ecom_metrics:
means = df[(df["Metric"] == col)]['Avg'].values
n_mean = np.interp(means, (mmin, mmax), (0, 1))
std = df[(df["Metric"] == col)]['Std'].values
n_std = np.interp(std, (smin, smax), (0, 1))
ax[1].plot(n_mean, label=legend_ecom[col])
# std = np.log10(df[(df["Metric"] == col)]['Std'].values)
# plot_mean_std(ax[1], n_mean, n_std, color=color3(col))
handles, labels = ax[1].get_legend_handles_labels()
lgd = ax[1].legend(handles, labels, loc='top left')
def plot_batch_results(df):
metrics = df['Metric'].unique()
health_metrics = [k for k in metrics if "Q" not in k and "Asym" not in k]
ecom_metrics = [k for k in metrics if "Q" in k]
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=[20, 5])
ax[0].set_title('Average Contagion Evolution')
ax[0].set_xlabel("Nº of Days")
ax[0].set_ylabel("% of Population")
for col in health_metrics:
means = df[(df["Metric"] == col)]['Avg'].values
std = df[(df["Metric"] == col)]['Std'].values
plot_mean_std(ax[0], means, std, legend=col, color=color1(col))
handles, labels = ax[0].get_legend_handles_labels()
lgd = ax[0].legend(handles, labels, loc='top right')
mmax = 0.0
mmin = np.inf
smax = 0
smin = np.inf
for col in ecom_metrics:
val = df[(df["Metric"] == col)]['Avg'].values
tmp = int(np.max(val))
mmax = np.max([mmax, tmp])
tmp = np.min(val)
mmin = np.min([mmin, tmp])
val = df[(df["Metric"] == col)]['Std'].values
tmp = np.max(val)
smax = np.max([smax, tmp])
tmp = np.min(val)
smin = np.min([smin, tmp])
ax[1].set_title('Average Economical Impact')
ax[1].set_xlabel("Nº of Days")
ax[1].set_ylabel("Wealth")
for col in ecom_metrics:
means = df[(df["Metric"] == col)]['Avg'].values
n_mean = np.interp(means, (mmin, mmax), (0, 1))
std = df[(df["Metric"] == col)]['Std'].values
n_std = np.interp(std, (smin, smax), (0, 1))
ax[1].plot(n_mean, label=legend_ecom[col])
#std = np.log10(df[(df["Metric"] == col)]['Std'].values)
#plot_mean_std(ax[1], n_mean, n_std, color=color3(col))
handles, labels = ax[1].get_legend_handles_labels()
lgd = ax[1].legend(handles, labels, loc='top left')
def plot_graph_batch_results(df, health_metrics=('Susceptible', 'Exposed','Infected', 'Hospitalization', 'Severe', 'Recovered_Immune', 'Death'),
ecom_metrics=('Q1', 'Q2', 'Q3', 'Q4', 'Q5','Business','Government'), **kwargs):
"""
Plot the results of a batch executions contained in the given DataFrame
:param ecom_metrics:
:param health_metrics:
:param df: Pandas DataFrame returned by batch_experiment method
"""
from matplotlib.ticker import (MultipleLocator, FormatStrFormatter,
AutoMinorLocator)
iterations = max(df['Iteration'].values) + 1
tick_unit = kwargs.get('tick_unit', 72)
tickslabels = [str(i // 24) for i in range(0, iterations, tick_unit)]
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=[20, 5])
ax[0].set_title('Average Contagion Evolution')
ax[0].set_xlabel("Nº of Days")
ax[0].set_ylabel("% of Population")
ax[0].set_xlim((0, iterations))
ax[0].xaxis.set_major_locator(MultipleLocator(tick_unit))
ax[0].set_xticklabels(tickslabels)
for col in health_metrics:
means = df[(df["Metric"] == col)]['Avg'].values
std = df[(df["Metric"] == col)]['Std'].values
plot_mean_std(ax[0], means, std, legend=col, color=color1(col))
handles, labels = ax[0].get_legend_handles_labels()
lgd = ax[0].legend(handles, labels, loc='top right')
ax[1].set_title('Average Economical Impact')
ax[1].set_xlabel("Nº of Days")
ax[1].set_ylabel("% of Wealth")
ax[1].set_xlim((0, iterations))
ax[1].xaxis.set_major_locator(MultipleLocator(tick_unit))
ax[1].set_xticklabels(tickslabels)
for col in ecom_metrics:
means = df[(df["Metric"] == col)]['Avg'].values
#n_mean = np.interp(means, (mmin, mmax), (0, 1))
std = df[(df["Metric"] == col)]['Std'].values
#n_std = np.interp(std, (smin, smax), (0, 1))
#ax[1].plot(means, label=legend_ecom[col])
# std = np.log10(df[(df["Metric"] == col)]['Std'].values)
plot_mean_std(ax[1], means, std, legend=legend_ecom[col], color=color3(col))
handles, labels = ax[1].get_legend_handles_labels()
lgd = ax[1].legend(handles, labels, loc='top left')
def plot_graph_batch_results(df, **kwargs):
"""
Plot the results of a batch executions contained in the given DataFrame
:param ecom_metrics:
:param health_metrics:
:param df: Pandas DataFrame returned by batch_experiment method
"""
from matplotlib.ticker import (MultipleLocator, FormatStrFormatter,
AutoMinorLocator)
A1 = ['Q1', 'Q2', 'Q3', 'Q4', 'Q5']
health_metrics = ('Susceptible', 'Infected', 'Hospitalization', 'Severe',
'Recovered_Immune', 'Death')
ecom_metrics = ('Q1', 'Q2', 'Q3', 'Q4', 'Q5', 'Business', 'Government')
health_legend = {
'Susceptible': '$S_t$ - Susceptible',
'Infected': '$I_t$ - Infected',
'Hospitalization': '$S^H_t$ - Hospitalization',
'Severe': '$I^S_t$ - Severe',
'Recovered_Immune': '$R_t$ - Recovered',
'Death': '$D_t$ - Death'
}
ecom_legend = {
'A1': '$W^{A1}_t$ - Persons',
'Business': '$W^{A3}_t$ - Business',
'Government': '$W^{A4}_t$ - Government',
}
colors = {
'A1': 'purple',
'Business': 'red',
'Government': 'brown',
}
epidem = kwargs.get('epidem', True)
iterations = max(df['Iteration'].values) + 1
tick_unit = kwargs.get('tick_unit', 72)
tickslabels = [str(i // 24) for i in range(0, iterations, tick_unit)]
fig, ax = plt.subplots(nrows=1,
ncols=1 if not epidem else 2,
figsize=[15, 4])
if epidem:
ep_ax = ax[0]
ep_ax.set_title('Average Epidemiological Evolution')
ep_ax.set_xlabel("Nº of Days")
ep_ax.set_ylabel("% of Population")
ep_ax.set_xlim((0, iterations))
ep_ax.xaxis.set_major_locator(MultipleLocator(tick_unit))
ep_ax.set_xticklabels(tickslabels)
for col in health_metrics:
means = df[(df["Metric"] == col)]['Avg'].values
std = df[(df["Metric"] == col)]['Std'].values
plot_mean_std(ep_ax,
means,
std,
legend=health_legend[col],
color=color1(col))
handles, labels = ep_ax.get_legend_handles_labels()
lgd = ep_ax.legend(handles, labels, loc=2, bbox_to_anchor=(1, 1))
ec_ax = ax[1] if epidem else ax
ec_ax.set_title('Average Economical Impact')
ec_ax.set_xlabel("Nº of Days")
ec_ax.set_ylabel("% of GDP")
ec_ax.set_xlim((0, iterations))
ec_ax.xaxis.set_major_locator(MultipleLocator(tick_unit))
ec_ax.set_xticklabels(tickslabels)
for col in ecom_legend.keys():
if col == 'A1':
means = np.zeros(iterations)
std = np.zeros(iterations)
for m2 in A1:
means += df[(df["Metric"] == m2)]['Avg'].values
std += df[(df["Metric"] == m2)]['Std'].values
else:
means = df[(df["Metric"] == col)]['Avg'].values
std = df[(df["Metric"] == col)]['Std'].values
l = len(means)
lb = [means[k] - std[k] for k in range(l)]
ub = [means[k] + std[k] for k in range(l)]
ec_ax.fill_between(range(l), ub, lb, color=colors[col], alpha=.4)
ec_ax.plot(range(l), means, colors[col], label=ecom_legend[col])
handles, labels = ec_ax.get_legend_handles_labels()
lgd = ec_ax.legend(handles, labels, loc=2, bbox_to_anchor=(1, 1))
plt.tight_layout()