def plot_mobility(series, label, stringency = None, until = None, annotation = "Google Mobility Data; baseline mobility measured from Jan 3 - Feb 6"):
plt.plot(series.date, smoothed(series.retail_and_recreation_percent_change_from_baseline), label = "Retail/Recreation")
plt.plot(series.date, smoothed(series.grocery_and_pharmacy_percent_change_from_baseline), label = "Grocery/Pharmacy")
plt.plot(series.date, smoothed(series.parks_percent_change_from_baseline), label = "Parks")
plt.plot(series.date, smoothed(series.transit_stations_percent_change_from_baseline), label = "Transit Stations")
plt.plot(series.date, smoothed(series.workplaces_percent_change_from_baseline), label = "Workplaces")
plt.plot(series.date, smoothed(series.residential_percent_change_from_baseline), label = "Residential")
if until:
right = pd.Timestamp(until)
elif stringency is not None:
right = stringency.Date.max()
else:
right = series.date.iloc[-1]
lax = plt.gca()
if stringency is not None:
plt.sca(lax.twinx())
stringency_IN = stringency.query("CountryName == 'India'")
stringency_US = stringency.query("(CountryName == 'United States') & (RegionName.isnull())", engine = "python")
plt.plot(stringency_IN.Date, stringency_IN.StringencyIndex, 'k--', alpha = 0.6, label = "IN Measure Stringency")
plt.plot(stringency_US.Date, stringency_US.StringencyIndex, 'k.' , alpha = 0.6, label = "US Measure Stringency")
plt.PlotDevice().ylabel("lockdown stringency index", rotation = -90, labelpad = 50)
plt.legend()
plt.sca(lax)
plt.legend(loc = "lower right")
plt.fill_betweenx((-100, 60), pd.to_datetime("March 24, 2020"), pd.to_datetime("June 1, 2020"), color = "black", alpha = 0.05, zorder = -1)
plt.text(s = "national lockdown", x = pd.to_datetime("April 27, 2020"), y = -90, fontdict = plt.theme.note, ha = "center", va = "top")
plt.PlotDevice()\
.xlabel("\ndate")\
.ylabel("% change in mobility\n")
# .title(f"\n{label}: Mobility & Lockdown Trends")\
# .annotate(annotation)\
plt.ylim(-100, 60)
plt.xlim(left = series.date.iloc[0], right = right)
def demand_curves(ranking_provider,
vax_policy,
phis=[25, 50, 100, 200],
phi_benchmark=25,
N_state=N_TN):
wtp_rankings = {phi: ranking_provider(phi, vax_policy) for phi in phis}
figure = plt.figure()
lines = []
# benchmark
benchmark = wtp_rankings[phi_benchmark]
x_pop = list(
chain(*zip(benchmark.loc[0]["num_vax"].shift(1).fillna(0),
benchmark.loc[0]["num_vax"])))
y_wtp = list(
chain(*zip(benchmark.loc[0]["wtp_pc_usd"], benchmark.loc[0]
["wtp_pc_usd"])))
lines.append(
plt.plot(x_pop, y_wtp, figure=figure, color="black", linewidth=2)[0])
lines.append(plt.plot(0, 0, color="white")[0])
# plot dynamic curve
for (phi, all_wtp) in wtp_rankings.items():
daily_doses = phi * percent * annually * N_state
distributed_doses = 0
x_pop = []
y_wtp = []
t_vax = []
ranking = 0
for t in range(simulation_range):
wtp = all_wtp.loc[t].reset_index()
ranking = wtp[(wtp.index >= ranking)
& (wtp.num_vax > distributed_doses)].index.min()
if np.isnan(ranking):
break
x_pop += [distributed_doses, distributed_doses + daily_doses]
t_vax += [t, t + 1]
y_wtp += [wtp.iloc[ranking].wtp_pc_usd] * 2
distributed_doses += daily_doses
lines.append(
plt.plot(x_pop,
y_wtp,
label=f"dynamic, {vax_policy}, $\phi = ${phi}%",
figure=figure)[0])
plt.legend(lines, [f"static, t = 0, $\phi = ${phi_benchmark}%", ""] +
[f"dynamic, {vax_policy}, $\phi = ${phi}%" for phi in phis],
title="allocation",
title_fontsize="24",
fontsize="20")
plt.xticks(fontsize="20")
plt.yticks(fontsize="20")
plt.PlotDevice().ylabel("WTP (USD)\n").xlabel("\nnumber vaccinated")
plt.ylim(0, 350)
plt.xlim(left=0, right=N_TN)
plt.show()
def plot_district_age_distribution(percentiles,
ylabel,
fmt,
phi=50,
vax_policy="random",
N_jk=None,
n=5,
district_spacing=1.5,
age_spacing=0.1,
rotation=0):
fig = plt.figure()
district_ordering = list(districts_to_run.index)[:n]
for (i, district) in enumerate(district_ordering):
ylls = percentiles[district, phi, vax_policy]
for j in range(7):
plt.errorbar(x=[district_spacing * i + age_spacing * (j - 3)],
y=ylls[1, 6 - j] * USD /
(N_jk[f"N_{6-j}"][district] if N_jk else 1),
yerr=[[(ylls[1, 6 - j] - ylls[0, 6 - j]) * USD /
(N_jk[f"N_{6-j}"][district] if N_jk else 1)],
[(ylls[2, 6 - j] - ylls[1, 6 - j]) * USD /
(N_jk[f"N_{6-j}"][district] if N_jk else 1)]],
fmt=fmt,
color=age_group_colors[6 - j],
figure=fig,
label=None if i > 0 else age_bin_labels[6 - j],
ms=12,
elinewidth=5)
plt.xticks([1.5 * _ for _ in range(n)],
district_ordering,
rotation=rotation,
fontsize="20")
plt.yticks(fontsize="20")
plt.legend(title="age bin",
title_fontsize="20",
fontsize="20",
ncol=7,
loc="lower center",
bbox_to_anchor=(0.5, 1))
ymin, ymax = plt.ylim()
plt.vlines(x=[0.75 + 1.5 * _ for _ in range(n - 1)],
ymin=ymin,
ymax=ymax,
color="gray",
alpha=0.5,
linewidths=2)
plt.ylim(ymin, ymax)
plt.gca().grid(False, axis="x")
plt.PlotDevice().title(f"\n{vax_policy} demand curves").ylabel(
f"{ylabel}\n")
def plot_state_age_distribution(percentiles,
ylabel,
fmt,
district_spacing=1.5,
n=5,
age_spacing=0.1,
rotation=0,
ymin=0,
ymax=1000):
fig = plt.figure()
state_ordering = list(
sorted(percentiles.keys(),
key=lambda k: percentiles[k][0].max(),
reverse=True))
for (i, state) in enumerate(state_ordering[:n]):
ylls = percentiles[state]
for j in range(7):
plt.errorbar(x=[district_spacing * i + age_spacing * (j - 3)],
y=ylls[0, 6 - j],
yerr=[[(ylls[0, 6 - j] - ylls[1, 6 - j])],
[(ylls[2, 6 - j] - ylls[0, 6 - j])]],
fmt=fmt,
color=agebin_colors[6 - j],
figure=fig,
label=None if i > 0 else agebin_labels[6 - j],
ms=12,
elinewidth=5)
plt.xticks([1.5 * _ for _ in range(n)],
state_ordering,
rotation=rotation,
fontsize="20")
plt.yticks(fontsize="20")
# plt.legend(title = "age bin", title_fontsize = "20", fontsize = "20", ncol = 7,
plt.legend(fontsize="20",
ncol=7,
loc="lower center",
bbox_to_anchor=(0.5, 1))
plt.vlines(x=[0.75 + 1.5 * _ for _ in range(n - 1)],
ymin=ymin,
ymax=ymax,
color="gray",
alpha=0.5,
linewidths=4)
plt.ylim(ymin, ymax)
plt.gca().grid(False, axis="x")
plt.PlotDevice().ylabel(f"{ylabel}\n")
def plot_component_breakdowns(color,
white,
colorlabel,
whitelabel,
semilogy=False,
ylabel="WTP (USD)"):
fig, ax = plt.subplots()
ax.bar(range(7),
white * USD,
bottom=color * USD,
color="white",
edgecolor=age_group_colors,
linewidth=2,
figure=fig)
ax.bar(range(7),
color * USD,
color=age_group_colors,
edgecolor=age_group_colors,
linewidth=2,
figure=fig)
ax.bar(range(7), [0],
label=whitelabel,
color="white",
edgecolor="black",
linewidth=2)
ax.bar(range(7), [0],
label=colorlabel,
color="black",
edgecolor="black",
linewidth=2)
plt.xticks(range(7), age_bin_labels, fontsize="20")
plt.yticks(fontsize="20")
plt.legend(ncol=4,
fontsize="20",
loc="lower center",
bbox_to_anchor=(0.5, 1))
plt.PlotDevice().ylabel(f"{ylabel}\n")
if semilogy: plt.semilogy()
def outcomes_per_policy(percentiles,
metric_label,
fmt,
phis=[25, 50, 100, 200],
reference=(25, "no_vax"),
reference_color=no_vax_color,
vax_policies=["contact", "random", "mortality"],
policy_colors=[
contactrate_vax_color, random_vax_color,
mortality_vax_color
],
policy_labels=[
"contact rate priority", "random assignment",
"mortality priority"
],
spacing=0.2):
fig = plt.figure()
md, lo, hi = percentiles[reference]
*_, bars = plt.errorbar(x=[0],
y=[md],
yerr=[[md - lo], [hi - md]],
figure=fig,
fmt=fmt,
color=reference_color,
label="no vaccination",
ms=12,
elinewidth=5)
[_.set_alpha(0.5) for _ in bars]
plt.hlines(md,
xmin=-1,
xmax=5,
linestyles="dotted",
colors=reference_color)
for (i, phi) in enumerate(phis, start=1):
for (j, (vax_policy, color, label)) in enumerate(
zip(vax_policies, policy_colors, policy_labels)):
md, lo, hi = death_percentiles[phi, vax_policy]
*_, bars = plt.errorbar(x=[i + spacing * (j - 1)],
y=[md],
yerr=[[md - lo], [hi - md]],
figure=fig,
fmt=fmt,
color=color,
label=label if i == 0 else None,
ms=12,
elinewidth=5)
[_.set_alpha(0.5) for _ in bars]
plt.legend(ncol=4,
fontsize="20",
loc="lower center",
bbox_to_anchor=(0.5, 1))
plt.xticks(range(len(phis) + 1),
[f"$\phi = {phi}$%" for phi in ([0] + phis)],
fontsize="20")
plt.yticks(fontsize="20")
plt.PlotDevice().ylabel(f"{metric_label}\n")
plt.gca().grid(False, axis="x")
ymin, ymax = plt.ylim()
plt.vlines(x=[0.5 + _ for _ in range(len(phis))],
ymin=ymin,
ymax=ymax,
color="gray",
alpha=0.5,
linewidths=2)
plt.ylim(ymin, ymax)
plt.xlim(-0.5, len(phis) + 1.5)
s=10,
zorder=10)
plot_TT, = plt.plot(idx,
dT_conf_scaled_smooth_TT[idx] / N_TT,
color=IN_color,
label="India (smoothed)",
figure=fig,
zorder=10,
linewidth=2)
plt.xticks(fontsize="20", rotation=0)
plt.yticks(fontsize="20")
plt.legend([scatter_TN, plot_TN, scatter_TT, plot_TT], [
"Tamil Nadu (raw)", "Tamil Nadu (smoothed)", "India (raw)",
"India (smoothed)"
],
fontsize="20",
ncol=4,
framealpha=1,
handlelength=0.75,
loc="lower center",
bbox_to_anchor=(0.5, 1))
plt.gca().xaxis.set_major_formatter(plt.bY_FMT)
plt.gca().xaxis.set_minor_formatter(plt.bY_FMT)
plt.xlim(left=pd.Timestamp("March 1, 2020"),
right=pd.Timestamp("April 15, 2021"))
plt.ylim(bottom=0)
plt.PlotDevice().ylabel("per-capita infection rate\n").xlabel("\ndate")
plt.show()
# 1B: per capita vaccination rates
vax = load_vax_data()\
.reindex(pd.date_range(start = pd.Timestamp("Jan 1, 2021"), end = simulation_start, freq = "D"), fill_value = 0)\
'Cumulative total per thousand': "total_per_thousand",
'Daily change in cumulative total per thousand': "delta_per_thousand",
'7-day smoothed daily change': "smoothed_delta",
'7-day smoothed daily change per thousand': "smoothed_delta_per_thousand",
'Short-term positive rate': "positivity",
'Short-term tests per case': "tests_per_case"
}
testing = pd.read_csv("data/covid-testing-all-observations.csv",
parse_dates=["Date"])
testing = testing[testing["ISO code"] == "IND"]\
.dropna()\
[schema.keys()]\
.rename(columns = schema)
testing["month"] = testing.date.dt.month
def formula(order: int) -> str:
powers = " + ".join(f"np.power(delta_per_thousand, {i + 1})"
for i in range(order))
return f"smoothed_delta ~ -1 + daily_tests + C(month)*({powers})"
model = OLS.from_formula(formula(order=3), data=testing).fit()
print(summary_col(model, regressor_order=["daily_tests"], drop_omitted=True))
plt.plot(0.2093 * df["TT"][:, "delta", "tested"], label="test-scaled")
plt.plot(df["TT"][:, "delta", "confirmed"], label="confirmed")
plt.legend()
plt.show()
label=None if i > 0 else [
"contact rate prioritized", "random assignment",
"mortality prioritized"
][dx],
ms=12,
elinewidth=5)
[_.set_alpha(0.5) for _ in bars]
plt.xticks(
range(len(metrics)),
[f"{phi}%" for phi in (100 * (10**np.linspace(-1, 1, 11))).round(0)],
fontsize="20")
plt.yticks(fontsize="20")
plt.PlotDevice().ylabel("deaths\n").xlabel(
"\npercentage of population vaccinated annually")
# plt.ylim(200, 450)
plt.legend(fontsize="20")
plt.show()
# evaluated_YLL_percentiles = {k: np.percentile(v, [5, 50, 95]) for (k, v) in evaluated_YLLs.items() if "ve70" in k or "novaccination" in k}
# contact_percentiles = {k: v for (k, v) in evaluated_YLL_percentiles.items() if "contact" in k}
# random_percentiles = {k: v for (k, v) in evaluated_YLL_percentiles.items() if "random" in k}
# mortality_percentiles = {k: v for (k, v) in evaluated_YLL_percentiles.items() if "mortality" in k}
# novax_percentiles = {k: v for (k, v) in evaluated_YLL_percentiles.items() if "novacc" in k}
# fig = plt.figure()
# plt.errorbar(
# x = [-1],
# y = novax_percentiles["novaccination"][1],
# yerr = [novax_percentiles["novaccination"][1] - [novax_percentiles["novaccination"][0]],[novax_percentiles["novaccination"][2] - novax_percentiles["novaccination"][1]]],
# fmt = "o",
# color = no_vax_color,
download_data(data, 'timeseries.json', "https://api.covid19india.org/v3/")
# data prep
with (data/'timeseries.json').open("rb") as fp:
df = flat_table.normalize(pd.read_json(fp)).fillna(0)
df.columns = df.columns.str.split('.', expand = True)
dates = np.squeeze(df["index"][None].values)
df = df.drop(columns = "index").set_index(dates).stack([1, 2]).drop("UN", axis = 1)
series = mobility[mobility.sub_region_1.isna()]
plt.plot(series.date, smoothed(series.retail_and_recreation_percent_change_from_baseline), label = "Retail/Recreation")
plt.fill_betweenx((-100, 60), pd.to_datetime("March 24, 2020"), pd.to_datetime("June 1, 2020"), color = "black", alpha = 0.05, zorder = -1)
plt.text(s = "national lockdown", x = pd.to_datetime("April 27, 2020"), y = -20, fontdict = plt.note_font, ha = "center", va = "top")
plt.ylim(-100, 10)
plt.xlim(series.date.min(), series.date.max())
plt.legend(loc = 'upper right')
lax = plt.gca()
plt.sca(lax.twinx())
plt.plot(df["TT"][:, "delta", "confirmed"].index, smoothed(df["TT"][:, "delta", "confirmed"].values), label = "Daily Cases", color = plt.PRED_PURPLE)
plt.legend(loc = 'lower right')
plt.PlotDevice().ylabel("new cases", rotation = -90, labelpad = 50)
plt.sca(lax)
plt.PlotDevice().title("\nIndia Mobility and Case Count Trends")\
.annotate("Google Mobility Data + Covid19India.org")\
.xlabel("\ndate")\
.ylabel("% change in mobility\n")
plt.show()
plt.plot(series.date, smoothed(series.retail_and_recreation_percent_change_from_baseline), label = "Retail/Recreation")
plt.fill_betweenx((-100, 60), pd.to_datetime("March 24, 2020"), pd.to_datetime("June 1, 2020"), color = "black", alpha = 0.05, zorder = -1)
plt.text(s = "national lockdown", x = pd.to_datetime("April 27, 2020"), y = -20, fontdict = plt.note_font, ha = "center", va = "top")
100 * (distributed_doses + daily_doses) / N_natl
]
t_vax += [t, t + 1]
y_tev += [tev.iloc[ranking].pc_tev_usd] * 2
distributed_doses += daily_doses
# lines += [plt.plot(x_pop, y_tev, label = f"dynamic, {vax_policy}, $\phi = ${phi}%", figure = figure)[0]]
lines += [
plt.plot(x_pop,
y_tev,
label=f"dynamic, {vax_policy}, $phi = ${phi}%",
figure=figure)[0]
]
plt.legend(
lines,
# ["static, t = 0, $\phi = $50%", ""] + [f"dynamic, {vax_policy}, $\phi = ${phi}%" for phi in phis],
["static, t = 0, $phi = $50%", ""] +
[f"dynamic, {vax_policy}, $phi = ${phi}%" for phi in phis],
title="allocation",
title_fontsize="24",
fontsize="20")
plt.xticks(fontsize="20")
plt.yticks(fontsize="20")
plt.PlotDevice().ylabel("TEV (USD)\n").xlabel(
"\npercentage of country vaccinated")
plt.ylim(0, 1600)
plt.xlim(left=0, right=100)
plt.show()
# 3C: YLL per million choropleth
if "3C" in figs_to_run or run_all:
india = gpd.read_file(data/"india.geojson")\
.drop(columns = ["id", "dt_code", "st_code", "year"])\
random_seed=0)
total_t = 0
schedule = [(1.01, 75), (1.4, 75), (0.9, 75)]
R0_timeseries = []
for (R0, t) in schedule:
R0_timeseries += [R0] * t
sir_model.Rt0 = R0
sir_model.run(t)
total_t += t
plt.plot(sir_model.dT)
plt.show()
plt.plot(R0_timeseries, "-", color="black", label="$R_0$")
plt.plot(sir_model.Rt, "-", color="dodgerblue", label="$R_t$")
plt.legend(framealpha=1, handlelength=1, loc="best")
plt.PlotDevice().xlabel("time").ylabel("reproductive rate").adjust(left=0.10,
bottom=0.15,
right=0.99,
top=0.99)
plt.ylim(0.5, 1.5)
plt.show()
# 1: parametric scheme:
dates, Rt, Rt_lb, Rt_ub, *_, anomalies, anomaly_dates = analytical_MPVS(
pd.DataFrame(sir_model.dT),
smoothing=convolution("uniform", 2),
CI=0.99,
totals=False)
pd = plt.Rt(dates, Rt, Rt_ub, Rt_lb, ymin = 0.5, ymax = 2.5, CI = 0.99, yaxis_colors = False, format_dates = False, critical_threshold = False)\
.xlabel("time")\
marker="s",
s=5,
alpha=0.5)
lineplot, = plt.plot(infections.index[:-7],
smoothed(infections.values[:-7]),
color="#CC4C75",
linewidth=2)
plt.PlotDevice()\
.l_title("daily confirmed cases in India")\
.r_title("source:\nCOVID19India")\
.axis_labels(x = "date", y = "cases", enforce_spacing = True)\
.adjust(left = 0.10, bottom = 0.15, right = 0.98, top = 0.90)
plt.xlim(infections.index[0] - one_day, infections.index[-1] + one_day)
plt.legend([scatter, lineplot],
["reported infection counts", "7-day moving average"],
handlelength=0.5,
framealpha=0,
prop={'size': 16})
plt.show()
# fig 2
mob2020 = pd.read_csv("data/2020_IN_Region_Mobility_Report.csv",
parse_dates=["date"])
mob2021 = pd.read_csv("data/2021_IN_Region_Mobility_Report.csv",
parse_dates=["date"])
mob = pd.concat([mob2020, mob2021]).set_index("date")
IN_mob = mob[mob.sub_region_1.isna()]\
.filter(like = "_percent_change_from_baseline", axis = 1).mean(axis = 1)
MH_mob = mob[(mob.sub_region_1 == "Maharashtra") & mob.sub_region_2.isna()]\
plt.figure()
plt.Rt(dates, Rt_pred, RR_CI_lower, RR_CI_upper, CI)\
.ylabel("Estimated $R_t$")\
.xlabel("Date")\
.title(district)\
.size(11, 8)\
.save(figs/f"Rt_est_MP{district}.png", dpi=600, bbox_inches="tight")#\
#.show()
plt.close()
from matplotlib.dates import DateFormatter
formatter = DateFormatter("%b\n%Y")
f = notched_smoothing(window=smoothing)
plt.plot(ts.loc["Maharashtra"].index,
ts.loc["Maharashtra"].Hospitalized,
color="black",
label="raw case counts from API")
plt.plot(ts.loc["Maharashtra"].index,
f(ts.loc["Maharashtra"].Hospitalized),
color="black",
linestyle="dashed",
alpha=0.5,
label="smoothed, seasonality-adjusted case counts")
plt.PlotDevice()\
.l_title("daily case counts in Maharashtra")\
.axis_labels(x = "date", y = "daily cases")
plt.gca().xaxis.set_major_formatter(formatter)
plt.legend(prop=plt.theme.note, handlelength=1, framealpha=0)
plt.show()
fmt="o",
color=clr,
label=None if i > 0 else [
"contact rate prioritized", "random assignment",
"mortality prioritized"
][dx],
ms=12,
elinewidth=5)
[_.set_alpha(0.5) for _ in bars]
plt.xticks(list(range(-1, len(metrics))),
[f"$\phi = {phi}$%" for phi in [0, 25, 50, 75, 100, 200, 400]],
fontsize="20")
plt.yticks(fontsize="20")
plt.PlotDevice().ylabel("\ndeaths")
plt.legend(fontsize="20", ncol=4, loc="lower center", bbox_to_anchor=(0.5, 1))
plt.show()
evaluated_YLL_percentiles = {
k: np.percentile(v, [5, 50, 95])
for (k, v) in evaluated_YLLs.items() if "ve70" in k or "novaccination" in k
}
contact_percentiles = {
k: v
for (k, v) in evaluated_YLL_percentiles.items() if "contact" in k
}
random_percentiles = {
k: v
for (k, v) in evaluated_YLL_percentiles.items() if "random" in k
}
mortality_percentiles = {
np.pad([np.mean(_) for _ in model.dT[1:]], (0, max_t - len(model.Rt)))
for model in models.values())
prob_deathx = dDx / Sx
prob_death = dD / S
sns.set_palette(age_group_colors)
dt = [simulation_start + pd.Timedelta(n, "days") for n in range(max_t - 1)]
PrD = pd.DataFrame(prob_deathx).T\
.rename(columns = dict(enumerate(IN_age_structure.keys())))\
.assign(t = dt)\
.set_index("t")
PrD.plot()
plt.legend(title="Age category",
title_fontsize=18,
fontsize=16,
framealpha=1,
handlelength=1)
plt.xlim(right=pd.Timestamp("Jan 01, 2022"))
plt.PlotDevice()\
.xlabel("\nDate")\
.ylabel("Probability\n")
plt.subplots_adjust(left=0.12, bottom=0.12, right=0.94, top=0.96)
plt.gca().xaxis.set_minor_locator(mpl.ticker.NullLocator())
plt.gca().xaxis.set_minor_formatter(mpl.ticker.NullFormatter())
plt.gca().xaxis.set_major_locator(mdates.AutoDateLocator())
plt.gca().xaxis.set_major_formatter(mdates.DateFormatter('%b %d'))
plt.xticks(fontsize="16")
plt.yticks(fontsize="16")
plt.gca().xaxis.grid(True, which="major")
plt.semilogy()