https://github.com/CSSEGISandData/COVID-19
"""
#Avoid re-reading case data if it's already stored in memory
try:
cases
except:
print("--> Reading in COVID-19 case data from Johns Hopkins CSSE")
if worldometers == True: include_repatriated = False
if read_from_local == True:
cases = pickle.load(open('cases_us.pickle','rb'))
dates = cases['dates']
del cases['dates']
else:
output = read_data.read_us(worldometers=worldometers)
dates = output['dates']
cases = output['cases']
#========================================================================================================
# Create plot based on type
#========================================================================================================
#Create figure
fig,ax = plt.subplots(figsize=(9,6),dpi=125)
#Total count
total_count = np.array([0.0 for i in cases['new york']['date']])
total_count_rp = np.array([0.0 for i in cases['new york']['date']])
#Iterate through every region
def plot_chart(plot_type=None):
#========================================================================================================
# User-defined settings
#========================================================================================================
#Whether to save image or display. "directory_path" string is ignored if setting==False.
save_image = {
'setting': True,
'directory_path': "full_directory_path_here"
}
#What to plot (confirmed, deaths, recovered, active, daily)
if plot_type == None:
plot_type = "active"
#Include repatriated cases (e.g., cruises)?
include_repatriated = False
#Plot total numbers?
plot_total = True
#Additional settings
settings = {
'log_y': True, #Use logarithmic y-axis?
'condensed_plot': True, #Condensed plot? (small dots and narrow lines)
'highlight_state': 'New York', #Highlight state?
'number_of_state': 20, #Limit number of states plotted?
}
#========================================================================================================
# Get COVID-19 case data
#========================================================================================================
repatriated_locations = ['diamond princess', 'grand princess']
"""
COVID-19 case data is retrieved from Johns Hopkins CSSE:
https://github.com/CSSEGISandData/COVID-19
"""
#Avoid re-reading case data if it's already stored in memory
try:
cases
except:
print("--> Reading in COVID-19 case data from Johns Hopkins CSSE")
output = read_data.read_us()
dates = output['dates']
cases = output['cases']
#========================================================================================================
# Create plot based on type
#========================================================================================================
max_value = 0
#Create figure
fig, ax = plt.subplots(figsize=(9, 6), dpi=125)
#Total count
total_count = np.array([0.0 for i in cases['diamond princess']['date']])
total_count_rp = np.array([0.0 for i in cases['diamond princess']['date']])
#Iterate through every region
sorted_keys = [
y[1] for y in sorted([(np.nanmax(cases[x][plot_type]), x)
for x in cases.keys()])
][::-1]
sorted_value = [
y[0] for y in sorted([(np.nanmax(cases[x][plot_type]), x)
for x in cases.keys()])
][::-1]
for idx, (key, value) in enumerate(zip(sorted_keys, sorted_value)):
#Special handling for Diamond Princess
if include_repatriated == False and key in repatriated_locations:
continue
#Total count
if key not in repatriated_locations:
total_count += np.array(cases[key][plot_type])
total_count_rp += np.array(cases[key][plot_type])
#Skip plotting if zero
if value == 0: continue
#How many states to plot?
lim = 19
if 'number_of_states' in settings.keys():
lim = settings['number_of_states'] - 1
if lim > 19: lim = 19
#Plot type
if idx > 19:
pass
else:
mtype = '--'
zord = 2
if np.nanmax(cases[key][plot_type]) > np.percentile(
sorted_value, 95):
mtype = '-o'
zord = 3
zord = 22 - idx
#Handle narrow plot
kwargs = {
'linewidth': 1.0,
'zorder': zord,
'label': f"{key.title()} ({cases[key][plot_type][-1]})"
}
if 'condensed_plot' in settings.keys(
) and settings['condensed_plot'] == True:
mtype = '-o'
kwargs['linewidth'] = 0.5
kwargs['ms'] = 2
#Highlight individual state
if 'highlight_state' in settings.keys(
) and settings['highlight_state'].lower() == key.lower():
kwargs['linewidth'] = 2.0
if 'ms' in kwargs.keys(): kwargs['ms'] = 4
#Plot lines
plt.plot(cases[key]['date'], cases[key][plot_type], mtype,
**kwargs)
max_value = max(max_value, max(cases[key][plot_type]))
#Plot total count
if plot_total == True:
kwargs = {
'zorder': 2,
'label': f'Total ({int(total_count[-1])})',
'color': 'k',
'linewidth': 2
}
plt.plot(cases[key]['date'], total_count, ':', **kwargs)
#plt.plot(cases[key]['date'],total_count,':',zorder=2,label=f'Total ({int(total_count[-1])})',color='k',linewidth=2)
if include_repatriated == True:
kwargs[
'label'] = f'Total (+ Repatriated)\n({int(total_count_rp[-1])})'
kwargs['color'] = 'b'
plt.plot(cases[key]['date'], total_count_rp, ':', **kwargs)
max_value = max(max_value, max(total_count))
#Format x-ticks
ax.set_xticks(cases[key]['date'][::7])
ax.set_xticklabels(cases[key]['date'][::7])
ax.xaxis.set_major_formatter(mdates.DateFormatter('%b\n%d'))
#Plot grid and legend
plt.grid()
kwargs = {
'loc': 2,
'title': f"Top {lim+1} locations plotted",
'prop': {
'size': 8
}
}
plt.legend(**kwargs).set_zorder(51)
#Plot title
title_string = {
'confirmed': 'Cumulative COVID-19 Confirmed Cases',
'deaths': 'Cumulative COVID-19 Deaths',
'recovered': 'Cumulative COVID-19 Recovered Cases',
'active': 'Daily COVID-19 Active Cases',
'daily': 'Daily COVID-19 New Cases',
}
add_title = "\n(Non-Repatriated Cases)" if include_repatriated == False else ""
plt.title(f"{title_string.get(plot_type)} {add_title}",
fontweight='bold',
loc='left')
plt.xlabel("Date", fontweight='bold')
plt.ylabel("Cases", fontweight='bold')
#Add logarithmic y-scale
if 'log_y' in settings.keys() and settings['log_y'] == True:
plt.yscale('log')
y_locs, y_labels = plt.yticks()
for i, loc in enumerate(y_locs):
if loc == 1.e+00: y_labels[i] = "1"
elif loc == 1.e+01: y_labels[i] = "10"
elif loc == 1.e+02: y_labels[i] = "100"
elif loc == 1.e+03: y_labels[i] = "1K"
elif loc == 1.e+04: y_labels[i] = "10K"
elif loc == 1.e+05: y_labels[i] = "100K"
elif loc == 1.e+06: y_labels[i] = "1M"
elif loc == 1.e+07: y_labels[i] = "10M"
elif loc == 1.e+08: y_labels[i] = "100M"
elif loc == 1.e+09: y_labels[i] = "1B"
elif loc == 1.e+10: y_labels[i] = "10B"
plt.yticks(y_locs, y_labels)
plt.ylim(bottom=1)
if max_value < 10: plt.ylim(top=10)
elif max_value < 100: plt.ylim(top=100)
elif max_value < 1000: plt.ylim(top=1000) #1K
elif max_value < 10000: plt.ylim(top=10000)
elif max_value < 100000: plt.ylim(top=100000)
elif max_value < 1000000: plt.ylim(top=1000000) #1M
elif max_value < 10000000: plt.ylim(top=10000000)
elif max_value < 100000000: plt.ylim(top=100000000)
elif max_value < 1000000000: plt.ylim(top=1000000000) #1B
elif max_value < 10000000000: plt.ylim(top=10000000000)
#Plot attribution
plt.title(
f'Data from Johns Hopkins CSSE{add_title}\nPlot by Stephen Mullens @srmullens\nCode adapted from Tomer Burg @burgwx',
loc='right',
fontsize=8)
if plot_type == "active":
kwargs = {
'fontweight': 'bold',
'ha': 'right',
'va': 'top',
'fontsize': 8
}
active_text = "\"Active\" cases = confirmed total - recovered - deaths"
plt.text(0.99, 0.99, active_text, transform=ax.transAxes, **kwargs)
#Show plot and close
if save_image['setting'] == True:
savepath = os.path.join(save_image['directory_path'],
f"{plot_type}_chart_us.png")
plt.savefig(savepath, bbox_inches='tight')
else:
plt.show()
plt.close()
#Alert script is done
print("Done!")
https://github.com/CSSEGISandData/COVID-19
"""
#Avoid re-reading case data if it's already stored in memory
try:
cases
except:
print("--> Reading in COVID-19 case data from Johns Hopkins CSSE")
if worldometers == True: include_repatriated = False
if read_from_local == True:
cases = pickle.load(open('cases_us.pickle', 'rb'))
dates = cases['dates']
del cases['dates']
else:
output = read_data.read_us(negative_daily=False,
worldometers=worldometers)
dates = output['dates']
cases = output['cases']
if plot_end_today == True: plot_end_date = dates[-1]
#========================================================================================================
# Create plot based on type
#========================================================================================================
#Function for returning number within range
def return_val(start_range, end_range, start_size, end_size, val):
frac = (val - start_range) / (end_range - start_range)
frac = (frac * (end_size - start_size)) + start_size
return frac
def plot_chart(plot_type=None):
#========================================================================================================
# User-defined settings
#========================================================================================================
#Whether to save image or display. "directory_path" string is ignored if setting==False.
save_image = {
'setting': True,
'directory_path': "full_directory_path_here"
}
#What to plot (confirmed, confirmed_normalized, deaths, recovered, active, daily)
if plot_type == None:
plot_type = "active"
#Include repatriated cases (e.g., cruises)?
include_repatriated = False
#Plot total numbers?
plot_total = True
#Over how many days should the doubling rate be calculated?
lag_days = 7
#Additional settings
settings = {
'log_y': True, #Use logarithmic y-axis?
'condensed_plot': True, #Condensed plot? (small dots and narrow lines)
'highlight_state': 'New York', #Highlight state?
'number_of_state': 20, #Limit number of states plotted?
}
#========================================================================================================
# Get COVID-19 case data
#========================================================================================================
#Z-Order:
# 2-highlighted trend
# 3-total trend
# 4-highlighted state dot
# 5-total dot
# 6-state dots
repatriated_locations = ['diamond princess', 'grand princess']
"""
COVID-19 case data is retrieved from Johns Hopkins CSSE:
https://github.com/CSSEGISandData/COVID-19
"""
abbr_state = read_data.abbr_state()
state_abbr = {v: k for k, v in abbr_state.items()}
#Avoid re-reading case data if it's already stored in memory
try:
cases
except:
print("--> Reading in COVID-19 case data from Johns Hopkins CSSE")
output = read_data.read_us()
dates = output['dates']
cases = output['cases']
#========================================================================================================
# Create plot based on type
#========================================================================================================
max_value = 0
max_doubling = -6
min_doubling = 6
lag_index = -lag_days - 1
highlighted_series = []
#Create figure
fig, ax = plt.subplots(figsize=(9, 6), dpi=125)
#Total count
total_count = np.array([0.0 for i in cases['diamond princess']['date']])
total_count_rp = np.array([0.0 for i in cases['diamond princess']['date']])
#Iterate through every region
sorted_keys = [
y[1] for y in sorted([(np.nanmax(cases[x][plot_type]), x)
for x in cases.keys()])
][::-1]
sorted_value = [
y[0] for y in sorted([(np.nanmax(cases[x][plot_type]), x)
for x in cases.keys()])
][::-1]
for idx, (key, value) in enumerate(zip(sorted_keys, sorted_value)):
end_day = cases[key][plot_type][-1]
start_day = cases[key][plot_type][lag_index]
day_after_start = cases[key][plot_type][lag_index + 1]
doubling_time = 999
inverse_doubling_time = 999
if end_day > 1:
if start_day > 0 and end_day != start_day and start_day != day_after_start:
doubling_time = lag_days * np.log(2) / np.log(
end_day / start_day)
inverse_doubling_time = 1 / doubling_time
#Special handling for Diamond Princess
if include_repatriated == False and key in repatriated_locations:
continue
#Total count
if key not in repatriated_locations:
total_count += np.array(cases[key][plot_type])
total_count_rp += np.array(cases[key][plot_type])
#Plot states, including highlighted state if applicable.
if inverse_doubling_time != 999:
# Plot highlighted state
if 'highlight_state' in settings.keys(
) and settings['highlight_state'].lower() == key.lower():
highlight_color = 'red'
highlight_key = f"{key.title()}"
highlight_doubling = f"{doubling_time:.1f}"
highlight_total = f"{end_day}"
kwargs = {'zorder': 4, 'color': highlight_color}
ax.scatter(inverse_doubling_time, end_day, **kwargs)
highlighted_series = cases[key][plot_type]
# Plot rest of states
else:
country_text_color = 'k'
ST = state_abbr.get(key.title())
if (ST == None): print(key.title(), ST)
kwargs = {
'zorder': 6,
'color': country_text_color,
'ha': 'center',
'va': 'center',
'family': 'monospace',
'fontsize': 8
}
ax.text(inverse_doubling_time, end_day, ST, **kwargs)
#Output stats to terminal
print(
f"{key.title()}\t{start_day}->{end_day}\t{doubling_time:.2f}")
#Store values useful for the plot axes.
max_value = max(max_value, end_day)
max_doubling = max(max_doubling, inverse_doubling_time)
min_doubling = min(min_doubling, inverse_doubling_time)
print(f"\nRange: {1/min_doubling:.2f} to {1/max_doubling:.2f} days")
#Calculate highlighted state running doubling time
if 'highlight_state' in settings.keys():
highlighted_series_doubling_time = []
length_hs = len(highlighted_series) - lag_days
for i in range(length_hs):
if highlighted_series[
i + lag_days] > 100 and highlighted_series[i] > 0:
highlighted_series_doubling_time.append(
1 / (lag_days * np.log(2) /
np.log(highlighted_series[i + lag_days] /
highlighted_series[i])))
#Plot line of highlighted series doubling time history
if len(highlighted_series_doubling_time) > 6:
kwargs = {
'zorder': 2,
'color': highlight_color,
'lw': 1,
'markevery': [-7],
'ms': 2
}
plt.plot(
highlighted_series_doubling_time,
highlighted_series[-len(highlighted_series_doubling_time):],
'-o', **kwargs)
hc_7 = True
elif len(highlighted_series_doubling_time) > 1:
kwargs = {'zorder': 2, 'color': highlight_color, 'lw': 1}
plt.plot(
highlighted_series_doubling_time,
highlighted_series[-len(highlighted_series_doubling_time):],
**kwargs)
hc_7 = False
else:
hc_7 = False
#Calculate US running doubling time
length = len(total_count) - lag_days
total_running_doubling_time = []
total_rp_running_doubling_time = []
for i in range(length):
if total_count[i + lag_days] > 100:
total_running_doubling_time.append(
1 / (lag_days * np.log(2) /
np.log(total_count[i + lag_days] / total_count[i])))
total_rp_running_doubling_time.append(
1 / (lag_days * np.log(2) /
np.log(total_count_rp[i + lag_days] / total_count_rp[i])))
#Plot total count
if plot_total == True:
total_doubling_time = lag_days * np.log(2) / np.log(
total_count[-1] / total_count[lag_index])
total_inverse_doubling_time = 1 / total_doubling_time
total_color = 'k'
total_doubling_title = f"{total_doubling_time:.1f}"
total_count_title = f"{int(total_count[-1])}"
kwargs = {'zorder': 5, 'color': total_color}
plt.scatter(total_inverse_doubling_time, total_count[-1], **kwargs)
kwargs = {
'zorder': 3,
'lw': 1,
'color': total_color,
'markevery': [-7],
'ms': 2
}
plt.plot(total_running_doubling_time,
total_count[-len(total_running_doubling_time):], "-o",
**kwargs)
#Store values useful for the plot.
max_value = max(max_value, total_count[-1])
max_doubling = max(
max_doubling,
1 / (1 * np.log(2) / np.log(total_count[-1] / total_count[-2])))
min_doubling = min(
min_doubling,
1 / (1 * np.log(2) / np.log(total_count[-1] / total_count[-2])))
#Plot grid and legend
plt.grid()
legend_title = f"Calculated from change\nbetween {cases[key]['date'][lag_index]:%b %d} and {cases[key]['date'][-1]:%b %d}"
#Blank entry for legend.
if 'highlight_state' in settings.keys():
plt.plot(
[], [],
'-o',
color=highlight_color,
label=
f"{highlight_key} & trend after 100 cases\n Time-{highlight_doubling} days, Total-{highlight_total}"
)
if plot_total == True:
plt.plot(
[], [],
'-o',
label=
f'US Total & trend\n Time-{total_doubling_title} days, Total-{int(total_count_title)}',
color=total_color)
ax.scatter([], [], color='white', label="Abbreviated States & Territories")
dot_l = False
if 'highlight_state' in settings.keys() and hc_7 == True:
p_hc = plt.scatter([], [], color=highlight_color, s=2)
if plot_total == True:
p_tc = plt.scatter([], [], color=total_color, s=2)
#print("hc,tc")
dot_h = [(p_hc, p_tc)]
dot_l = ["Small dots were 7 days ago"]
else:
#print("hc")
dot_h = [(p_hc)]
dot_l = ["Small dot was 7 days ago"]
else:
if plot_total == True:
p_tc = plt.scatter([], [], color=total_color, s=2)
#print("tc")
dot_h = [(p_tc)]
dot_l = ["Small dot was 7 days ago"]
handles, labels = ax.get_legend_handles_labels()
if dot_l:
handles = handles + dot_h
labels = labels + dot_l
kwargs = {'loc': 2, 'prop': {'size': 8}}
if plot_type == 'deaths': kwargs['loc'] = 1
plt.legend(handles,
labels,
title=legend_title,
handler_map={
tuple: HandlerTuple(ndivide=None)
},
**kwargs).set_zorder(51)
#Format x-ticks
xticks = [
1 / -0.5, 1 / -1, 1 / -2, 1 / -3, 1 / -4, 1 / -5, 1 / -6, 1 / -7, 0,
1 / 7, 1 / 6, 1 / 5, 1 / 4, 1 / 3, 1 / 2, 1 / 1, 1 / 0.5
]
xtick_labels = [
"Halving\nEvery\nHalf-day", "Halving\nEvery\nDay",
"Halving\nEvery\n2 Days", "Halving\nEvery\n3 Days",
"Halving\nEvery\n4 Days", "Halving\nEvery\n5 Days",
"Halving\nEvery\n6 Days", "Halving\nEvery\nWeek", "no\nchange",
"Doubling\nEvery\nWeek", "Doubling\nEvery\n6 Days",
"Doubling\nEvery\n5 Days", "Doubling\nEvery\n4 Days",
"Doubling\nEvery\n3 Days", "Doubling\nEvery\n2 Days",
"Doubling\nEvery\nDay", "Doubling\nEvery\nHalf-day"
]
#Format x-axis
if np.absolute(max_doubling) > np.absolute(min_doubling):
left = -max_doubling - 0.1
right = max_doubling + 0.1
else:
left = min_doubling - 0.1
right = -min_doubling + 0.1
if plot_type != 'active': left = 0
if right < 0.333 or (left == 0 and right < 0.5):
xticks = xticks[:5] + xticks[7:10] + xticks[12:]
xtick_labels = xtick_labels[:5] + xtick_labels[7:10] + xtick_labels[12:]
elif right < 0.5 or (left == 0 and right < 1):
xticks = xticks[:4] + xticks[7:10] + xticks[13:]
xtick_labels = xtick_labels[:4] + xtick_labels[7:10] + xtick_labels[13:]
elif right < 1 or (left == 0 and right < 2):
xticks = xticks[:3] + xticks[7:10] + xticks[14:]
xtick_labels = xtick_labels[:3] + xtick_labels[7:10] + xtick_labels[14:]
elif right < 2 or (left == 0 and right >= 2):
xticks = xticks[:3] + xticks[7:10] + xticks[14:]
xtick_labels = xtick_labels[:3] + [
xtick_labels[7], "", xtick_labels[9]
] + xtick_labels[14:]
else:
xticks = xticks[:3] + xticks[8] + xticks[6:]
xtick_labels = xtick_labels[:3] + [""] + xtick_labels[6:]
ax.set_xticks(xticks)
ax.set_xticklabels(xtick_labels)
plt.xlim(left, right)
#Add logarithmic y-scale
if 'log_y' in settings.keys() and settings['log_y'] == True:
plt.yscale('log')
y_locs, y_labels = plt.yticks()
for i, loc in enumerate(y_locs):
if loc == 1.e+00: y_labels[i] = "1"
elif loc == 1.e+01: y_labels[i] = "10"
elif loc == 1.e+02: y_labels[i] = "100"
elif loc == 1.e+03: y_labels[i] = "1K"
elif loc == 1.e+04: y_labels[i] = "10K"
elif loc == 1.e+05: y_labels[i] = "100K"
elif loc == 1.e+06: y_labels[i] = "1M"
elif loc == 1.e+07: y_labels[i] = "10M"
elif loc == 1.e+08: y_labels[i] = "100M"
elif loc == 1.e+09: y_labels[i] = "1B"
elif loc == 1.e+10: y_labels[i] = "10B"
plt.yticks(y_locs, y_labels)
plt.ylim(bottom=1)
if max_value < 10: plt.ylim(top=10)
elif max_value < 100: plt.ylim(top=100)
elif max_value < 1000: plt.ylim(top=1000) #1K
elif max_value < 10000: plt.ylim(top=10000)
elif max_value < 100000: plt.ylim(top=100000)
elif max_value < 1000000: plt.ylim(top=1000000) #1M
elif max_value < 10000000: plt.ylim(top=10000000)
elif max_value < 100000000: plt.ylim(top=100000000)
elif max_value < 1000000000: plt.ylim(top=1000000000) #1B
elif max_value < 10000000000: plt.ylim(top=10000000000)
#Plot title
title_string = {
'confirmed': 'Doubling Time of Cumulative COVID-19 Confirmed Cases',
'deaths': 'Doubling Time of Cumulative COVID-19 Deaths',
'recovered': 'Doubling Time of Cumulative COVID-19 Recovered Cases',
'active': 'Doubling Time of Daily COVID-19 Active Cases',
'daily': 'Doubling Time of Daily COVID-19 New Cases',
}
add_title = "\n(Non-Repatriated Cases)" if include_repatriated == False else ""
plt.title(f"{title_string.get(plot_type)} {add_title}",
fontweight='bold',
loc='left')
if left == 0:
xlabel_text = "<--slower increase\t\t\tfaster increase-->".expandtabs()
else:
xlabel_text = "<--faster decrease\t\t\tfaster increase-->".expandtabs()
plt.xlabel(xlabel_text, fontweight='bold')
plt.ylabel("Number of Cases", fontweight='bold')
#Plot attribution
plt.title(
f'Data from Johns Hopkins CSSE\nPlot by Stephen Mullens @srmullens\nCode adapted from Tomer Burg @burgwx',
loc='right',
fontsize=6)
if plot_type == "active":
kwargs = {
'fontweight': 'bold',
'ha': 'right',
'va': 'top',
'fontsize': 8
}
active_text = "\"Active\" cases = confirmed total - recovered - deaths"
plt.text(0.99, 0.99, active_text, transform=ax.transAxes, **kwargs)
#Show plot and close
if save_image['setting'] == True:
savepath = os.path.join(save_image['directory_path'],
f"{plot_type}_doubling_us.png")
plt.savefig(savepath, bbox_inches='tight')
else:
plt.show()
plt.close()
#Alert script is done
print("Done!")
dates = cases['dates']
del cases['dates']
else:
output = read_data.read_world(negative_daily=False,
worldometers=worldometers)
dates = output['dates']
cases = output['cases']
if plot_end_today == True: plot_end_date = dates[-1]
#Substitute US for states
if us_states == True:
del cases['us']
for case in cases.keys():
cases[case]['us'] = False
output_us = read_data.read_us(negative_daily=False)
cases_us = output_us['cases']
for case in cases_us.keys():
cases_us[case]['us'] = True
cases.update(cases_us)
#========================================================================================================
# Create plot based on type
#========================================================================================================
#Function for returning number within range
def return_val(start_range, end_range, start_size, end_size, val):
frac = (val - start_range) / (end_range - start_range)
frac = (frac * (end_size - start_size)) + start_size
return frac