def estimate(time_series: pd.Series) -> pd.DataFrame:
estimates = analytical_MPVS(time_series,
CI=CI,
smoothing=smooth,
totals=True)
return pd.DataFrame(
data={
"date": estimates[0],
"Rt": estimates[1],
"Rt_upper": estimates[2],
"Rt_lower": estimates[3],
"total_cases": estimates[-4][2:],
"new_cases": estimates[-3],
})
def run_adaptive_model(df: pd.DataFrame, CI: float, smoothing: Callable,
filepath: Path) -> None:
'''
Runs adaptive control model of Rt and smoothed case counts based on what is currently in the
analytical_MPVS module. Takes in dataframe of cases and saves to csv a dataframe of results.
'''
# Initialize results df
res_full = pd.DataFrame()
# Loop through each state
print("Estimating state Rt values...")
for state in tqdm(df['state'].unique()):
# Calculate Rt for that state
state_df = df[df['state'] == state].set_index('date')
(dates, RR_pred, RR_CI_upper, RR_CI_lower, T_pred, T_CI_upper,
T_CI_lower, total_cases, new_cases_ts, _,
anomaly_dates) = analytical_MPVS(
state_df[state_df['positive'] > 0]['positive'],
CI=CI,
smoothing=smoothing)
assert (len(dates) == len(RR_pred))
# Save results
res = pd.DataFrame({
'state': state,
'date': dates,
'RR_pred': RR_pred,
'RR_CI_upper': RR_CI_upper,
'RR_CI_lower': RR_CI_lower,
'T_pred': T_pred,
'T_CI_upper': T_CI_upper,
'T_CI_lower': T_CI_lower,
'new_cases_ts': new_cases_ts,
'total_cases': total_cases[2:],
'anamoly': dates.isin(set(anomaly_dates))
})
res_full = pd.concat([res_full, res], axis=0)
# Merge results back onto input df and return
merged_df = df.merge(res_full, how='outer', on=['state', 'date'])
merged_df.to_csv(filepath / "adaptive_estimates.csv")
smoothing = 10
CI = 0.95
state_cases = pd.read_csv(data / "Bihar_cases_data_Oct03.csv",
parse_dates=["date_reported"],
dayfirst=True)
state_ts = state_cases["date_reported"].value_counts().sort_index()
district_names, population_counts, _ = etl.district_migration_matrix(
data / "Migration Matrix - District.csv")
populations = dict(zip(district_names, population_counts))
# first, look at state level predictions
(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(
state_ts,
CI=CI,
smoothing=notched_smoothing(window=smoothing),
totals=False)
plt.Rt(dates, Rt_pred[1:], Rt_CI_upper[1:], Rt_CI_lower[1:], CI, ymin=0, ymax=4)\
.title("\nBihar: Reproductive Number Estimate")\
.annotate(f"data from {str(dates[0]).split()[0]} to {str(dates[-1]).split()[0]}")\
.xlabel("date")\
.ylabel("$R_t$", rotation=0, labelpad=20)\
.show()
np.random.seed(33)
Bihar = SIR("Bihar",
99_000_000,
dT0=T_pred[-1],
Rt0=Rt_pred[-1],
# model details
CI = 0.95
smoothing = 30
alpha = 3.8
beta = 2.25
vs = 0.999
true_Rt = pd.read_table("./true_Rt.txt", dtype="float", squeeze=True)
obs_cases = pd.read_table("./obs_cases.txt", dtype="float", squeeze=True)
(dates, RR_pred, RR_CI_upper, RR_CI_lower, T_pred, T_CI_upper, T_CI_lower,
total_cases, new_cases_ts, anomalies, anomaly_dates) = analytical_MPVS(
obs_cases,
CI=CI,
alpha=alpha,
beta=beta,
variance_shift=vs,
smoothing=lambda ts: box_filter(ts, smoothing, smoothing // 2))
print("Rt today:", RR_pred[-1])
print("a, b, vs, MSE:", alpha, beta, vs,
((true_Rt.loc[len(true_Rt) - len(RR_pred):] - RR_pred)**2).sum())
plot_RR_est(dates, RR_pred, RR_CI_lower, RR_CI_upper, CI)\
.ylabel("Estimated $R_t$")\
.title("Synthetic Data Estimation")\
.size(11, 8)
plt.plot(true_Rt.index, true_Rt.values, 'k--', label="True $R_t$")
plt.xlim(0, 150)
plt.ylim(0, 2.5)
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",
"confirmed"],
CI=CI,
smoothing=smooth,
totals=False)
(dates_D, Rt_D, Rtu_D, Rtl_D, *_) = analytical_MPVS(df[state][:, "delta",
"deceased"],
CI=CI,
smoothing=smooth,
totals=False)
plt.Rt(dates_I, Rt_I, Rtu_I, Rtl_I, CI)\
.title(f"{state} - $R_t(I)$ estimator")
plt.figure()
plt.Rt(dates_D, Rt_D, Rtu_D, Rtl_D, CI)\
.title(f"{state} - $R_t(D)$ estimator")
plt.show()
return None
cases = pd.read_csv("data/1 Nop 2020 Data collection template update South Sulawesi_update (01112020) (2).csv", usecols = schema.keys())\
.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))
cases.age = cases.age.apply(parse_age)
cases = cases.query("regency == 'Makassar'").dropna(subset = ["age"])
cases["age_bin"] = pd.cut(cases.age, [0, 20, 100], labels = ["school", "nonschool"])
cases = cases[cases.confirmed <= "Oct 25, 2020"]
age_ts = cases[["age_bin", "confirmed"]].groupby(["age_bin", "confirmed"]).size().sort_index()
(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(age_ts.loc["school"], CI = CI, smoothing = smoothing, totals = False)
school_Rt = np.mean(Rt_pred[-14:])
school_T_lb = T_CI_lower[-1]
school_T_ub = T_CI_upper[-1]
plt.Rt(dates, Rt_pred, Rt_CI_upper, Rt_CI_lower, CI)\
.title("\nMakassar: Reproductive Number Estimate: school-age population")\
.xlabel("\ndate")\
.ylabel("$R_t$\n", rotation=0, labelpad=30)\
.annotate(f"\n{window}-day smoothing window, gamma-prior Bayesian estimation method")\
.show()
(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(age_ts.loc["nonschool"], CI = CI, smoothing = smoothing, totals = False)
run_date = str(pd.Timestamp.now()).split()[0]
ts = get_time_series(df, "detected_state")
states = ["Maharashtra", "Bihar", "Delhi", "Andhra Pradesh", "Telangana", "Tamil Nadu", "Madhya Pradesh"]
for state in states:
print(state)
print(" + running estimation...")
(
dates,
Rt_pred, RR_CI_upper, RR_CI_lower,
T_pred, T_CI_upper, T_CI_lower,
total_cases, new_cases_ts,
anomalies, anomaly_dates
) = analytical_MPVS(ts.loc[state].Hospitalized, CI = CI, smoothing = notched_smoothing(window = smoothing), totals = False)
estimates = pd.DataFrame(data = {
"dates": dates,
"Rt_pred": Rt_pred,
"RR_CI_upper": RR_CI_upper,
"RR_CI_lower": RR_CI_lower,
"T_pred": T_pred,
"T_CI_upper": T_CI_upper,
"T_CI_lower": T_CI_lower,
"total_cases": total_cases[2:],
"new_cases_ts": new_cases_ts,
})
print(" + Rt today:", Rt_pred[-1])
# plt.Rt(dates, Rt_pred, RR_CI_lower, RR_CI_upper, CI)\
# .ylabel("Estimated $R_t$")\
figs = root / "figs/comparison/kaggle"
states = [
"Maharashtra"
] #, "Bihar", "Delhi", "Andhra Pradesh", "Telangana", "Tamil Nadu", "Madhya Pradesh"]
kaggle = pd.read_csv(data / "covid_19_india.csv",
parse_dates=[1],
dayfirst=True).set_index("Date")
for state in states:
print(state)
print(" + running estimation...")
(dates, RR_pred, RR_CI_upper, RR_CI_lower, T_pred, T_CI_upper, T_CI_lower,
total_cases, new_cases_ts, anomalies, anomaly_dates) = analytical_MPVS(
kaggle[kaggle["State/UnionTerritory"] == state].Confirmed,
CI=CI,
smoothing=lambda ts: box_filter(ts, smoothing, 3))
estimates = pd.DataFrame(
data={
"dates": dates,
"RR_pred": RR_pred,
"RR_CI_upper": RR_CI_upper,
"RR_CI_lower": RR_CI_lower,
"T_pred": T_pred,
"T_CI_upper": T_CI_upper,
"T_CI_lower": T_CI_lower,
"total_cases": total_cases[2:],
"new_cases_ts": new_cases_ts,
})
print(" + Rt today:", RR_pred[-1])
(data, figs) = setup(level = "INFO")
# for province in provinces:
# logger.info("downloading data for %s", province)
# download_data(data, filename(province), base_url = "https://data.covid19.go.id/public/api/")
province_cases = {province: load_province_timeseries(data, province) for province in provinces}
bgn = min(cases.index.min() for cases in province_cases.values())
end = max(cases.index.max() for cases in province_cases.values())
idx = pd.date_range(bgn, end)
province_cases = {province: cases.reindex(idx, method = "pad").fillna(0) for (province, cases) in province_cases.items()}
natl_cases = sum(province_cases.values())
logger.info("running national-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(natl_cases, CI = CI, smoothing = smoothing)
plt.Rt(dates, Rt_pred, Rt_CI_upper, Rt_CI_lower, CI, ymin=0, ymax=4)\
.title("\nIndonesia: Reproductive Number Estimate")\
.xlabel("\ndate")\
.ylabel("$R_t$", rotation=0, labelpad=30)\
.annotate(f"\n{window}-day smoothing window, gamma-prior Bayesian estimation method")\
.show()
logger.info("running case-forward prediction")
IDN = SIR("IDN", 267.7e6, dT0 = T_pred[-1], Rt0 = Rt_pred[-1], mobility = 0, random_seed = 0).run(14)
logger.info("province-level projections")
migration = np.zeros((len(provinces), len(provinces)))
estimates = []
province: load_province_timeseries(data, province, "Apr 1, 2020")
for province in provinces
}
bgn = min(cases.index.min() for cases in province_cases.values())
end = max(cases.index.max() for cases in province_cases.values())
idx = pd.date_range(bgn, end)
province_cases = {
province: cases.reindex(idx, method="pad").fillna(0)
for (province, cases) in province_cases.items()
}
prediction_period = 14 * days
for province in provinces:
title = province.title().replace("Dki", "DKI")
logger.info(title)
(dates, RR_pred, RR_CI_upper, RR_CI_lower, T_pred, T_CI_upper, T_CI_lower, total_cases, new_cases_ts, anomalies, anomaly_dates)\
= analytical_MPVS(province_cases[province], CI = CI, smoothing = smoothing)
IDN = Model.single_unit(name = province, population = priority_pops[province], I0 = T_pred[-1], RR0 = RR_pred[-1], upper_CI = T_CI_upper[-1], lower_CI = T_CI_lower[-1], mobility = 0, random_seed = 0)\
.run(prediction_period)
plt.Rt(dates, RR_pred, RR_CI_upper, RR_CI_lower, CI, ymin=0.2, ymax=4.5)\
.title(f"{title}")\
.xlabel("\ndate")\
.ylabel("$R_t$", rotation=0, labelpad=30)\
.show()
# plt.daily_cases(dates, T_pred, T_CI_upper, T_CI_lower, new_cases_ts, anomaly_dates, anomalies, CI, IDN[0].delta_T[:-1], IDN[0].lower_CI[1:], IDN[0].upper_CI[1:])\
# .title(f"\n{title}")\
# .xlabel("\ndate")\
# .ylabel("cases")\
# .show()
.set_axis(dkij.columns.str.lower(), 1)\
.assign(
district = dkij.district.str.title(),
subdistrict = dkij.subdistrict.apply(lambda name: next((k for (k, v) in replacements.items() if name in v), name)),
)
district_cases = dkij.groupby(["district", "date_positiveresult"])["id"].count().sort_index()
districts = sorted(dkij.district.unique())
migration = np.zeros((len(districts), len(districts)))
R_mandatory = dict()
R_voluntary = dict()
max_len = 1 + max(map(len, districts))
with tqdm(districts) as progress:
for district in districts:
progress.set_description(f"{district :<{max_len}}")
(dates, RR_pred, *_) = analytical_MPVS(district_cases.loc[district], CI = CI, smoothing = notched_smoothing(window = window), totals=False)
Rt = pd.DataFrame(data = {"Rt": RR_pred[1:]}, index = dates)
R_mandatory[district] = np.mean(Rt[(Rt.index > "April 1, 2020") & (Rt.index < "June 1, 2020")])[0]
R_voluntary[district] = np.mean(Rt[(Rt.index < "April 1, 2020")])[0]
pops = [
2_430_410,
910_381,
2_164_070,
2_817_994,
1_729_444,
23_011
]
gdf = gpd.read_file("data/gadm36_IDN_shp/gadm36_IDN_2.shp")\
.query("NAME_1 == 'Jakarta Raya'")\
cases = pd.read_stata(data/"coviddkijakarta_290920.dta")\
.query("province == 'DKI JAKARTA'")\
.drop(columns = dkij_drop_cols + ["province"])
cases = cases\
.set_axis(cases.columns.str.lower(), 1)\
.assign(
district = cases.district.str.title(),
subdistrict = cases.subdistrict.apply(lambda name: next((k for (k, v) in replacements.items() if name in v), name)),
)
cases["age_bin"] = pd.cut(cases.age,
bins=[0] + list(range(20, 80, 10)) + [100])
age_ts = cases[["age_bin", "date_positiveresult"
]].groupby(["age_bin",
"date_positiveresult"]).size().sort_index()
dkij_max_rts = {}
(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(age_ts.sum(level = 1), CI = CI, smoothing = smoothing, totals = False)
r = pd.Series(Rt_pred, index=dates)
dkij_max_rts["all"] = r[r.index.month_name() == "April"].max()
for age_bin in age_ts.index.get_level_values(0).categories:
print(age_bin)
(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(age_ts.loc[age_bin], CI = CI, smoothing = smoothing, totals = False)
r = pd.Series(Rt_pred, index=dates)
dkij_max_rts[age_bin] = r[r.index.month_name() == "April"].max()
print(dkij_max_rts)
def estimate(ts, smoothing):
(state_dates, R, *_) = analytical_MPVS(ts.Hospitalized, smoothing = smoothing)
dates = [sd[1] if isinstance(sd, tuple) else sd for sd in state_dates]
return pd.DataFrame({"date": dates, "R": R}).set_index("date")
df = load_all_data(
v3_paths = [data/filepath for filepath in paths['v3']],
v4_paths = [data/filepath for filepath in paths['v4']]
)
data_recency = str(df["date_announced"].max()).split()[0]
run_date = str(pd.Timestamp.now()).split()[0]
ts = get_time_series(df[df.detected_state == "Delhi"])
(
dates,
RR_pred, RR_CI_upper, RR_CI_lower,
T_pred, T_CI_upper, T_CI_lower,
total_cases, new_cases_ts,
anomalies, anomaly_dates
) = analytical_MPVS(ts.delta[ts.delta > 0], CI = CI, smoothing = convolution(window = smoothing))
#= analytical_MPVS(ts.Hospitalized[ts.Hospitalized > 0], CI = CI, smoothing = lambda ts: box_filter(ts, smoothing, 10))
np.random.seed(33)
delhi = Model([ModelUnit("Delhi", 18_000_000, I0 = T_pred[-1], RR0 = RR_pred[-1], mobility = 0)])
delhi.run(14, np.zeros((1,1)))
t_pred = [dates[-1] + pd.Timedelta(days = i) for i in range(len(delhi[0].delta_T))]
plot_RR_est(dates, RR_pred, RR_CI_upper, RR_CI_lower, CI)
PlotDevice().title("Delhi: Reproductive Number Estimate").xlabel("Date").ylabel("Rt", rotation=0, labelpad=20)
plt.show()
delhi[0].lower_CI[0] = T_CI_lower[-1]
delhi[0].upper_CI[0] = T_CI_upper[-1]
print(delhi[0].delta_T)
# plt.daily_cases(dates, T_pred, T_CI_upper, T_CI_lower, new_cases_ts, anomaly_dates, anomalies, CI,
# prediction_ts = [
# (current[0].delta_T[1:], current[0].lower_CI[1:], current[0].upper_CI[1:], "orange", r"projection with current $R_t$"),
# (target[0].delta_T[1:], target[0].lower_CI[1:], target[0].upper_CI[1:], "green", r"projection with $R_t \rightarrow 0.9$")
# ])\
# .adjust(left = 0.10, right = 0.95, bottom = 0.15, top = 0.95)\
# .xlabel("date")\
# .ylabel("cases")\
# .show()
# run Indian states
for (state, pop) in [("Maharashtra", 112374333), ("Gujarat", 60439692),
("Bihar", 104099452)]:
print(state)
(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(india[india.state == state][["date", "confirmed"]].set_index("date") , CI = CI, smoothing = smoothing, totals=True)
dates = [pd.Timestamp(_).to_pydatetime().date() for _ in dates]
anomaly_dates = [
pd.Timestamp(_).to_pydatetime().date() for _ in anomaly_dates
]
model = lambda: Model.single_unit(name=state,
RR0=Rt_pred[-1],
population=pop,
infectious_period=infectious_period,
I0=T_pred[-1],
lower_CI=T_CI_lower[-1],
upper_CI=T_CI_upper[-1],
random_seed=33)
forward_pred_period = 9
t_pred = [
dates[-1] + pd.Timedelta(days=i)
time_series["delta_I"] = time_series.groupby(level=0)['Hospitalized'].diff()
time_series["dow"] = time_series.index.get_level_values(1).dayofweek
plot_average_change(time_series,
"(All India)",
filename=figs / "avg_delta_I_DoW_India.png")
for state in tqdm(time_series.index.get_level_values(0).unique()):
plot_average_change(time_series.loc[state],
f"({state})",
filename=figs / f"avg_delta_I_DoW_{state}.png")
# are anomalies falling on certain days?
print("checking anomalies...")
smoothing = 5
(*_,
anomaly_dates) = analytical_MPVS(natl_time_series["Hospitalized"].iloc[:-1],
CI=0.95,
smoothing=convolution(window=smoothing))
anomaly_histogram(anomaly_dates,
"(All India)",
filename=figs / "anomaly_DoW_hist_India.png")
for state in tqdm(time_series.index.get_level_values(0).unique()):
(*_, anomaly_dates) = analytical_MPVS(
time_series.loc[state]["Hospitalized"].iloc[:-1],
CI=0.95,
smoothing=convolution(window=smoothing))
anomaly_histogram(anomaly_dates,
f"({state})",
filename=figs / f"anomaly_DoW_hist_{state}.png")
print("estimating spectral densities...")
# what does the aggregate spectral density look like?
for _ in top_level[timeseries]],
columns=["date", "total_cases"])
df["date"] = (date_scale * df["date"]).apply(pd.Timestamp)
return df.set_index("date")
logger.info("district-level projections")
pops = [sum([2_430_410, 910_381, 2_164_070, 2_817_994, 1_729_444, 23_011])]
dkij = load_province_timeseries(data, district)
R_mandatory = dict()
R_voluntary = dict()
(dates, Rt_pred,
*_) = analytical_MPVS(dkij,
CI=CI,
smoothing=notched_smoothing(window=window),
totals=True)
Rt = pd.DataFrame(data={"Rt": Rt_pred}, index=dates)
R_mandatory[district] = np.mean(Rt[(Rt.index >= "Sept 21, 2020")])[0]
R_voluntary[district] = np.mean(Rt[(Rt.index < "April 1, 2020")])[0]
si, sf = 0, 10
simulation_results = [
run_policies([dkij.iloc[-1][0] - dkij.iloc[-2][0]],
pops,
districts,
np.zeros((1, 1)),
gamma,
R_mandatory,
R_voluntary,
CI = 0.95
# private data
state_cases = pd.read_csv(data / "Bihar_cases_data_Jul23.csv",
parse_dates=["date_reported"],
dayfirst=True)
state_ts = state_cases["date_reported"].value_counts().sort_index()
district_names, population_counts, _ = etl.district_migration_matrix(
data / "Migration Matrix - District.csv")
populations = dict(zip(district_names, population_counts))
# first, look at state level predictions
(dates, RR_pred, RR_CI_upper, RR_CI_lower, T_pred, T_CI_upper, T_CI_lower,
total_cases, new_cases_ts, anomalies,
anomaly_dates) = analytical_MPVS(state_ts,
CI=CI,
smoothing=convolution(window=smoothing))
plot_RR_est(dates, RR_pred, RR_CI_upper, RR_CI_lower, CI, ymin=0, ymax=4)\
.title("Bihar: Reproductive Number Estimate Comparisons")\
.xlabel("Date")\
.ylabel("Rt", rotation=0, labelpad=20)
plt.ylim(0, 4)
# public data
paths = {
"v3": [data_path(_) for _ in (1, 2)],
"v4": [data_path(_) for _ in range(3, 13)]
}
for target in paths['v3'] + paths['v4']:
.query("province == 'DKI JAKARTA'")\
.drop(columns=dkij_drop_cols + ["province"])
dkij["district"] = dkij.district.str.title()
gdf = gpd.read_file("data/gadm36_IDN_shp/gadm36_IDN_2.shp")\
.query("NAME_1 == 'Jakarta Raya'")\
.drop(columns=shp_drop_cols)
bbox = shapely.geometry.box(minx = 106.65, maxx = 107.00, miny = -6.40, maxy=-6.05)
gdf = gdf[gdf.intersects(bbox)]
jakarta_districts = dkij.district.str.title().unique()
jakarta_cases = dkij.groupby("date_positiveresult")["id"].count().rename("cases")
logger.info("running province-level Rt estimate")
(dates, RR_pred, RR_CI_upper, RR_CI_lower, T_pred, T_CI_upper, T_CI_lower, total_cases, new_cases_ts, anomalies, anomaly_dates)\
= analytical_MPVS(jakarta_cases, CI = CI, smoothing = smoothing, totals=False)
plt.Rt(dates, RR_pred[1:], RR_CI_upper[1:], RR_CI_lower[1:], CI)\
.title("\nDKI Jakarta: 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")
prediction_period = 14*days
IDN = SIR(name = "IDN", population = 267.7e6, dT0 = T_pred[-1], Rt0 = RR_pred[-1], upper_CI = T_CI_upper[-1], lower_CI = T_CI_lower[-1], mobility = 0, random_seed = 0)\
.run(prediction_period)
plt.daily_cases(dates, T_pred[1:], T_CI_upper[1:], T_CI_lower[1:], new_cases_ts[1:], anomaly_dates, anomalies, CI,
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")\
.ylabel("reproductive rate")\
.adjust(left = 0.11, bottom = 0.15, right = 0.98, top = 0.98)
plt.plot(sir_model.Rt, "-", color="white", linewidth=3, zorder=10)
sim_rt, = plt.plot(sir_model.Rt,
"-",
color="dodgerblue",
linewidth=2,
zorder=11)
anoms = plt.vlines(anomaly_dates, 0, 4, colors="red", linewidth=2, alpha=0.5)
plt.legend([pd.markers["Rt"], sim_rt, anoms],
["Estimated $R_t$ (99% CI)", "simulated $R_t$", "anomalies"],
gamma = 0.2
smoothing = 10
CI = 0.95
download_data(data, 'state_wise_daily.csv')
state_df = load_statewise_data(data / "state_wise_daily.csv")
country_time_series = get_time_series(state_df)
estimates = []
timeseries = []
# country level
(dates, RR_pred, RR_CI_upper, RR_CI_lower, T_pred, T_CI_upper, T_CI_lower,
total_cases, new_cases_ts, anomalies, anomaly_dates) = analytical_MPVS(
country_time_series["Hospitalized"].iloc[:-1],
CI=CI,
smoothing=notched_smoothing(window=smoothing))
country_code = state_name_lookup["India"]
for row in zip(dates, RR_pred, RR_CI_upper, RR_CI_lower):
timeseries.append((country_code, *row))
# state level rt estimates
state_time_series = get_time_series(state_df, 'state')
state_names = list(state_time_series.index.get_level_values(level=0).unique())
max_len = 1 + max(map(len, state_names))
with tqdm(state_names) as states:
for state in states:
state_code = state_name_lookup[state]
states.set_description(f"{state :<{max_len}}")
try:
# .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")
prediction_period = 14*days
I0 = (~cases.confirmed.isna()).sum() - (~cases.recovered.isna()).sum() - (~cases.died.isna()).sum()
IDN = SIR(name = "IDN", population = 8_819_500, dT0 = T_pred[-1], Rt0 = Rt_pred[-1], upper_CI = T_CI_upper[-1], lower_CI = T_CI_lower[-1], mobility = 0, random_seed = 0, I0 = I0)\
.run(prediction_period)
plt.daily_cases(dates, T_pred, T_CI_upper, T_CI_lower, new_cases_ts, anomaly_dates, anomalies, CI,
age_structure_norm = sum(IN_age_structure.values())
IN_age_ratios = np.array(
[v / age_structure_norm for (k, v) in IN_age_structure.items()])
split_by_age = lambda v: (v * IN_age_ratios).astype(int)
# get age-specific prevalence from KA sero
KA = pd.read_stata("data/ka_cases_deaths_time_newagecat.dta")
KA.agecat = KA.agecat.where(
KA.agecat != 85,
75) # we don't have econ data for 85+ so combine 75+ and 85+ categories
KA_agecases = KA.groupby(["agecat", "date"])["patientcode"]\
.count().sort_index().rename("cases")\
.unstack().fillna(0).stack()
KA_ts = KA_agecases.sum(level=1)
(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(KA_ts, notched_smoothing(5))
COVID_age_ratios = (KA_agecases.sum(level=0) / KA_agecases.sum()).values
split_by_prevalence = lambda v: (v * IN_age_ratios).astype(int)
for seed in range(10):
model = AgeStructured("KA", 6.11e7, 857, 915345, 1.826,
diag(u) @ C, IN_age_ratios, COVID_age_ratios, seed)
counter = 0
while model.dT[-1].sum() > 0:
model.forward_epi_step()
counter += 1
print(seed, counter, model.dT)
if vax_pct_annual_goal == 0 and vax_effectiveness != 1.00:
continue
# grab time series
D, R = ts.loc[district][["dD", "dR"]].sum()
dT_conf_district = ts.loc[district].dT
dT_conf_district = dT_conf_district.reindex(pd.date_range(
dT_conf_district.index.min(), dT_conf_district.index.max()),
fill_value=0)
dT_conf_district_smooth = pd.Series(
smooth(dT_conf_district),
index=dT_conf_district.index).clip(0).astype(int)
# run Rt estimation on scaled timeseries
(Rt_dates, Rt_est, *_) = analytical_MPVS(T_ratio * dT_conf_district_smooth,
CI=CI,
smoothing=lambda _: _,
totals=False)
Rt = dict(zip(Rt_dates, Rt_est))
daily_rate = vax_pct_annual_goal / 365
daily_vax_doses = int(vax_effectiveness * daily_rate * N_district)
T_scaled = dT_conf_district_smooth.cumsum()[simulation_start] * T_ratio
model = SIR(
name=state,
population=N_district,
dT0=np.ones(num_sims) *
(dT_conf_district_smooth[simulation_start] * T_ratio).astype(int),
Rt0=Rt[simulation_start] * N_district / (N_district - T_scaled),
I0=np.ones(num_sims) * (T_scaled - R - D),
.query("age.str.strip() != ''", engine = "python")
parse_datetimes(cases.loc[:, "confirmed"])
cases.regency = cases.regency.str.title().map(
lambda s: regency_names.get(s, s))
cases.age = cases.age.apply(parse_age)
cases = cases.dropna(subset=["age"])
cases["age_bin"] = pd.cut(cases.age,
bins=[0] + list(range(20, 80, 10)) + [100])
age_ts = cases[["age_bin",
"confirmed"]].groupby(["age_bin",
"confirmed"]).size().sort_index()
ss_max_rts = {}
fig, axs = plt.subplots(4, 2, True, True)
(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(age_ts.sum(level = 1), CI = CI, smoothing = notched_smoothing(window = 5), totals = False)
plt.sca(axs.flat[0])
plt.Rt(dates, Rt_pred, Rt_CI_upper, Rt_CI_lower,
CI).annotate(f"all ages").adjust(left=0.04,
right=0.96,
top=0.95,
bottom=0.05,
hspace=0.3,
wspace=0.15)
r = pd.Series(Rt_pred, index=dates)
ss_max_rts["all"] = r[r.index.month_name() == "April"].max()
for (age_bin,
ax) in zip(age_ts.index.get_level_values(0).categories, axs.flat[1:]):
print(age_bin)
(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)\
gamma = 0.2
window = 10
CI = 0.95
state_cases = pd.read_csv(data/"Bihar_cases_data_Oct03.csv", parse_dates=["date_reported", "date_status_change"], dayfirst=True)
state_cases["geo_reported"] = state_cases.geo_reported.str.strip()
state_cases = state_cases[state_cases.date_reported <= "2020-09-30"]
state_ts = state_cases["date_reported"].value_counts().sort_index()
district_ts = state_cases.groupby(["geo_reported", "date_reported"])["date_reported"].count().sort_index()
districts, pops, migrations = etl.district_migration_matrix(data/"Migration Matrix - District.csv")
districts = sorted([etl.replacements.get(dn, dn) for dn in districts])
R_mandatory = dict()
for district in districts:#district_ts.index.get_level_values(0).unique():
try:
(_, Rt, *_) = analytical_MPVS(district_ts.loc[district], CI = CI, smoothing = notched_smoothing(window = 10), totals = False)
Rm = np.mean(Rt)
except ValueError as v:
Rm = 1.5
R_mandatory[district] = Rm
R_voluntary = {district: 1.2*R for (district, R) in R_mandatory.items()}
si, sf = 0, 10
simulation_results = [
run_policies(state_cases, pops, districts, migrations, gamma, R_mandatory, R_voluntary, lockdown_period = lockdown_period, total = total_time, seed = seed)
for seed in tqdm(range(si, sf))
]
plt.simulations(
ts_full = get_time_series(df, "detected_state")
ts = ts_full.query("status_change_date <= 'October 14, 2020'")
states = ["Bihar", "Maharashtra", "Punjab", "Tamil Nadu"][-1:]
codes = ["BR", "MH", "PN", "TN"][-1:]
pops = [99.02e6, 114.2e6, 27.98e6, 67.86e6][-1:]
Rt_ranges = [(0.9, 2.4), (0.55, 2.06), (0.55, 2.22), (0.84, 1.06)][-1:]
windows = [7, 14, 7, 10][-1:]
for (state, code, pop, Rt_range, smoothing) in zip(states, codes, pops, Rt_ranges, windows):
print(state)
print(" + running estimation...")
state_ts_full = pd.Series(data = notched_smoothing(window = smoothing)(ts_full.loc[state].Hospitalized), index = ts_full.loc[state].Hospitalized.index)
(dates, Rt_pred, RR_CI_upper, RR_CI_lower, T_pred, T_CI_upper, T_CI_lower, total_cases, new_cases_ts, anomalies, anomaly_dates)\
= analytical_MPVS(ts.loc[state].Hospitalized, CI = CI, smoothing = lambda x:x, totals = False)
Rt = pd.DataFrame({"Rt": Rt_pred}, index = dates)
Rt_m = np.mean(Rt[(Rt.index >= "31 March, 2020") & (Rt.index <= "17 May, 2020")])[0]
Rt_v = np.mean(Rt[(Rt.index < "31 March, 2020")])[0]
print(" + Rt today:", Rt_pred[-1])
print(" + Rt_m :", Rt_m)
print(" + Rt_v :", Rt_v)
historical = pd.DataFrame({"smoothed": new_cases_ts}, index = dates)
plt.Rt(dates, Rt_pred, RR_CI_lower, RR_CI_upper, CI)\
.ylabel("$R_t$")\
.xlabel("date")\
.title(f"\n{state}: Reproductive Number Estimate")\
.annotate(f"public data from {str(dates[0]).split()[0]} to {str(dates[-1]).split()[0]}")\
.show()
