# select order by minimizing AIC where coefficient on number of tests > 0
models = [
OLS.from_formula(scaling(order), data=obs).fit() for order in range(1, 10)
]
(model_idx, selected_model) = min(
((i, each)
for (i, each) in enumerate(models) if each.params["tested"] > 0),
key=lambda _: _[1].aic)
print(" i aic r2 beta")
for (i, model) in enumerate(models):
print("*" if i == model_idx else " ", i + 1, model.aic.round(2),
model.rsquared.round(2), model.params["tested"].round(2))
scale_factor = selected_model.params["tested"]
plt.plot(0.2093 * df[state][:, "delta", "tested"],
label="national test-scaled")
plt.plot(scale_factor * df[state][:, "delta", "tested"],
label="state test-scaled")
plt.plot(df[state][:, "delta", "confirmed"], label="confirmed")
plt.legend()
plt.PlotDevice().title(f"\n{state} / case scaling comparison").xlabel(
"\ndate").ylabel("cases\n")
plt.show()
# I vs D estimators
gamma = 0.2
window = 7 * days
CI = 0.95
smooth = notched_smoothing(window)
(dates_I, Rt_I, Rtu_I, Rtl_I, *_) = analytical_MPVS(df[state][:, "delta",
# plot simulation
plt.scatter(dT_conf["April 1, 2020":simulation_start].index,
dT_conf["April 1, 2020":simulation_start].values * T_ratio,
label="seroprevalence-scaled cases (pre-simulation)",
color="black",
s=5)
# plt.scatter(dT_conf[simulation_start:].index, dT_conf[simulation_start:].values*T_ratio, color = "grey", label = "seroprevalence-scaled cases (post-simulation)", s = 5)
# t = pd.Timestamp(date)
dates = pd.date_range(
simulation_start,
simulation_start + pd.Timedelta(len(model.dT) - 1, "days"))
# dates = pd.date_range(t, pd.Timestamp("April 1, 2021"))
n = len(dates)
plt.plot(dates,
np.array([_.mean().astype(int) for _ in model.dT][:n]),
label="mean simulated daily cases",
color="rebeccapurple")
plt.fill_between(dates, [_.min().astype(int) for _ in model.dT][:n],
[_.max().astype(int) for _ in model.dT][:n],
label="simulation range",
alpha=0.3,
color="rebeccapurple")
plt.vlines(pd.Timestamp(date),
1,
1e6,
linestyles="dashed",
label="date of seroprevalence study")
plt.legend(handlelength=1, framealpha=1)
plt.semilogy()
plt.xlim(pd.Timestamp("April 1, 2020"), dates[-1])
plt.ylim(1, 1e6)
.rename(columns = schema)\
.dropna(how = 'all')
parse_datetimes(cases.loc[:, "confirmed"])
cases.regency = cases.regency.str.title().map(lambda s: regency_names.get(s, s))
# generation_interval = cases[~cases.symptom_onset.isna() & ~cases.confirmed.isna()]\
# .apply(get_generation_interval, axis = 1)\
# .dropna()\
# .value_counts()\
# .sort_index()
# generation_interval = generation_interval[(generation_interval.index >= 0) & (generation_interval.index <= 60)]
# generation_interval /= generation_interval.sum()
new_cases = cases.confirmed.value_counts().sort_index()
new_cases_smoothed = smoothing(new_cases)
plt.plot(new_cases, '.', color = "blue")
plt.plot(new_cases.index, new_cases_smoothed, '-', color = "black")
plt.show()
logger.info("running province-level Rt estimate")
(dates, Rt_pred, Rt_CI_upper, Rt_CI_lower, T_pred, T_CI_upper, T_CI_lower, total_cases, new_cases_ts, anomalies, anomaly_dates)\
= analytical_MPVS(new_cases, CI = CI, smoothing = smoothing, totals = False)
plt.Rt(dates, Rt_pred, Rt_CI_upper, Rt_CI_lower, CI)\
.title("\nSouth Sulawesi: Reproductive Number Estimate")\
.xlabel("\ndate")\
.ylabel("$R_t$\n", rotation=0, labelpad=30)\
.annotate(f"\n{window}-day smoothing window, gamma-prior Bayesian estimation method")\
.show()
logger.info("running case-forward prediction")
population=500000,
I0=100,
dT0=20,
Rt0=1.01,
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),
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="upper left")
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.note_font,
ha="center",
va="top")
plt.PlotDevice()\
.title(f"\n{label}: Mobility & Lockdown Trends")\
.annotate(annotation)\
.xlabel("\ndate")\
.ylabel("% change in mobility\n")
plt.ylim(-100, 60)
plt.xlim(left=series.date.iloc[0], right=right)
from adaptive.etl.commons import download_data
from pathlib import Path
data = Path("./data")
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())
valid_idx = ~df.isna() & df.str.endswith("20")
valid = df[valid_idx]
monthfirst_idx = valid.str.endswith("/20") # short years -> month first notation
valid.loc[( monthfirst_idx)] = pd.to_datetime(valid[( monthfirst_idx)], errors = 'coerce', format = "%m/%d/%y", dayfirst = False)
valid.loc[(~monthfirst_idx)] = pd.to_datetime(valid[(~monthfirst_idx)], errors = 'coerce', format = "%d/%m/%Y", dayfirst = True)
# assert df.max() <= pd.to_datetime("October 03, 2020"), "date parsing resulted in future dates"
df.loc[valid_idx] = valid.apply(pd.Timestamp)
sulsel = pd.read_csv("data/3 OCT 2020 Data collection template update South Sulawesi_CASE.csv", usecols = schema.keys())\
.rename(columns = schema)\
.dropna(how = 'all')
parse_datetimes(sulsel.loc[:, "confirmed"])
sulsel = sulsel.confirmed.value_counts().sort_index()
plt.plot(dkij.index, dkij.values, color = "royalblue", label = "private")
plt.plot(dkij_public.diff(), color = "firebrick", label = "public")
plt.legend()
plt.PlotDevice()\
.title("\nJakarta: public vs private case counts")\
.xlabel("date")\
.ylabel("cases")
plt.xlim(right = dkij.index.max())
plt.ylim(top = 800)
plt.show()
plt.plot(sulsel, color = "royalblue", label = "private", linewidth = 3)
plt.plot(sulsel_public.diff(), color = "firebrick", label = "public")
plt.legend()
plt.PlotDevice()\
.title("\nSouth Sulawesi: public vs private case counts")\
schema = {
'Date': "date",
'Daily change in cumulative total': "daily_tests",
'Cumulative total': "total_tests",
'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()
