def load_observations(fips=None, ref_date=REF_DATE):
"""
Load observations (new cases, new deaths and hospitalizations) for
given fips code.
Parameters
----------
fips: str
FIPS code.
ref_date: Datetime
Reference start date.
Returns
-------
observations: pd.DataFrame
Contains observations for given fips codes, with columns:
- new_cases: float, observed new cases
- new_deaths: float, observed new deaths
- hospitalizations: float, observed hospitalizatons
and dates of observation as index.
"""
observations = {}
if len(fips) == 5:
times, observations['new_cases'], observations['new_deaths'] = \
load_data.load_new_case_data_by_fips(fips, ref_date)
hospital_times, hospitalizations, hospitalization_data_type = \
load_data.load_hospitalization_data(fips, t0=ref_date)
observations['times'] = times.values
elif len(fips) == 2:
state_obj = us.states.lookup(fips)
observations['times'], observations['new_cases'], observations['new_deaths'] = \
load_data.load_new_case_data_by_state(state_obj.name, ref_date)
hospital_times, hospitalizations, hospitalization_data_type = \
load_data.load_hospitalization_data_by_state(state_obj.abbr, t0=ref_date)
observations['times'] = np.array(observations['times'])
observations['hospitalizations'] = np.full(
observations['times'].shape[0], np.nan)
if hospitalization_data_type is HospitalizationDataType.CUMULATIVE_HOSPITALIZATIONS:
observations['hospitalizations'][
hospital_times -
observations['times'].min()] = np.diff(hospitalizations)
elif hospitalization_data_type is HospitalizationDataType.CURRENT_HOSPITALIZATIONS:
observations['hospitalizations'][
hospital_times -
observations['times'].min()] = hospitalizations
observation_dates = [
ref_date + timedelta(int(t)) for t in observations['times']
]
observations = pd.DataFrame(
observations,
index=pd.DatetimeIndex(observation_dates)).dropna(axis=1,
how='all')
return observations
def __init__(self,
fips,
ref_date=datetime(year=2020, month=1, day=1),
min_deaths=2,
n_years=1,
cases_to_deaths_err_factor=.5,
hospital_to_deaths_err_factor=.5,
percent_error_on_max_observation=0.5,
with_age_structure=False):
# Seed the random state. It is unclear whether this propagates to the
# Minuit optimizer.
np.random.seed(seed=42)
self.fips = fips
self.ref_date = ref_date
self.min_deaths = min_deaths
self.t_list = np.linspace(0, int(365 * n_years),
int(365 * n_years) + 1)
self.cases_to_deaths_err_factor = cases_to_deaths_err_factor
self.hospital_to_deaths_err_factor = hospital_to_deaths_err_factor
self.percent_error_on_max_observation = percent_error_on_max_observation
self.t0_guess = 60
self.with_age_structure = with_age_structure
if len(fips) == 2: # State FIPS are 2 digits
self.agg_level = AggregationLevel.STATE
self.state_obj = us.states.lookup(self.fips)
self.state = self.state_obj.name
self.times, self.observed_new_cases, self.observed_new_deaths = \
load_data.load_new_case_data_by_state(self.state, self.ref_date)
self.hospital_times, self.hospitalizations, self.hospitalization_data_type = \
load_data.load_hospitalization_data_by_state(self.state_obj.abbr, t0=self.ref_date)
self.display_name = self.state
else:
self.agg_level = AggregationLevel.COUNTY
geo_metadata = load_data.load_county_metadata().set_index(
'fips').loc[fips].to_dict()
state = geo_metadata['state']
self.state_obj = us.states.lookup(state)
county = geo_metadata['county']
if county:
self.display_name = county + ', ' + state
else:
self.display_name = state
# TODO Swap for new data source.
self.times, self.observed_new_cases, self.observed_new_deaths = \
load_data.load_new_case_data_by_fips(self.fips, t0=self.ref_date)
self.hospital_times, self.hospitalizations, self.hospitalization_data_type = \
load_data.load_hospitalization_data(self.fips, t0=self.ref_date)
self.cases_stdev, self.hosp_stdev, self.deaths_stdev = self.calculate_observation_errors(
)
self.set_inference_parameters()
self.model_fit_keys = ['R0', 'eps', 't_break', 'log10_I_initial']
self.SEIR_kwargs = self.get_average_seir_parameters()
self.fit_results = None
self.mle_model = None
self.chi2_deaths = None
self.chi2_cases = None
self.chi2_hosp = None
self.dof_deaths = None
self.dof_cases = None
self.dof_hosp = None
def __init__(
self,
fips,
window_size=InferRtConstants.COUNT_SMOOTHING_WINDOW_SIZE,
kernel_std=5,
r_list=np.linspace(0, 10, 501),
process_sigma=0.05,
ref_date=datetime(year=2020, month=1, day=1),
confidence_intervals=(0.68, 0.95),
min_cases=5,
min_deaths=5,
include_testing_correction=True,
):
np.random.seed(InferRtConstants.RNG_SEED)
# Param Generation used for Xcor in align_time_series, has some stochastic FFT elements.
self.fips = fips
self.r_list = r_list
self.window_size = window_size
self.kernel_std = kernel_std
self.process_sigma = process_sigma
self.ref_date = ref_date
self.confidence_intervals = confidence_intervals
self.min_cases = min_cases
self.min_deaths = min_deaths
self.include_testing_correction = include_testing_correction
# Because rounding is disabled we don't need high min_deaths, min_cases anymore
self.min_cases = min(InferRtConstants.MIN_COUNTS_TO_INFER, self.min_cases)
if not InferRtConstants.DISABLE_DEATHS:
self.min_deaths = min(InferRtConstants.MIN_COUNTS_TO_INFER, self.min_deaths)
if len(fips) == 2: # State FIPS are 2 digits
self.agg_level = AggregationLevel.STATE
self.state_obj = us.states.lookup(self.fips)
self.state = self.state_obj.name
(
self.times,
self.observed_new_cases,
self.observed_new_deaths,
) = load_data.load_new_case_data_by_state(
self.state,
self.ref_date,
include_testing_correction=self.include_testing_correction,
)
self.times_raw_new_cases, self.raw_new_cases, _ = load_data.load_new_case_data_by_state(
self.state, self.ref_date, include_testing_correction=False
)
(
self.hospital_times,
self.hospitalizations,
self.hospitalization_data_type,
) = load_data.load_hospitalization_data_by_state(
state=self.state_obj.abbr, t0=self.ref_date
)
self.display_name = self.state
else:
self.agg_level = AggregationLevel.COUNTY
self.geo_metadata = (
load_data.load_county_metadata().set_index("fips").loc[fips].to_dict()
)
self.state = self.geo_metadata["state"]
self.state_obj = us.states.lookup(self.state)
self.county = self.geo_metadata["county"]
if self.county:
self.display_name = self.county + ", " + self.state
else:
self.display_name = self.state
(
self.times,
self.observed_new_cases,
self.observed_new_deaths,
) = load_data.load_new_case_data_by_fips(
self.fips,
t0=self.ref_date,
include_testing_correction=self.include_testing_correction,
)
(
self.times_raw_new_cases,
self.raw_new_cases,
_,
) = load_data.load_new_case_data_by_fips(
self.fips, t0=self.ref_date, include_testing_correction=False,
)
(
self.hospital_times,
self.hospitalizations,
self.hospitalization_data_type,
) = load_data.load_hospitalization_data(self.fips, t0=self.ref_date)
self.case_dates = [ref_date + timedelta(days=int(t)) for t in self.times]
self.raw_new_case_dates = [
ref_date + timedelta(days=int(t)) for t in self.times_raw_new_cases
]
if self.hospitalization_data_type:
self.hospital_dates = [ref_date + timedelta(days=int(t)) for t in self.hospital_times]
self.default_parameters = ParameterEnsembleGenerator(
fips=self.fips, N_samples=500, t_list=np.linspace(0, 365, 366)
).get_average_seir_parameters()
# Serial period = Incubation + 0.5 * Infections
self.serial_period = (
1 / self.default_parameters["sigma"] + 0.5 * 1 / self.default_parameters["delta"]
)
# If we only receive current hospitalizations, we need to account for
# the outflow to reconstruct new admissions.
if (
self.hospitalization_data_type
is load_data.HospitalizationDataType.CURRENT_HOSPITALIZATIONS
):
los_general = self.default_parameters["hospitalization_length_of_stay_general"]
los_icu = self.default_parameters["hospitalization_length_of_stay_icu"]
hosp_rate_general = self.default_parameters["hospitalization_rate_general"]
hosp_rate_icu = self.default_parameters["hospitalization_rate_icu"]
icu_rate = hosp_rate_icu / hosp_rate_general
flow_out_of_hosp = self.hospitalizations[:-1] * (
(1 - icu_rate) / los_general + icu_rate / los_icu
)
# We are attempting to reconstruct the cumulative hospitalizations.
self.hospitalizations = np.diff(self.hospitalizations) + flow_out_of_hosp
self.hospital_dates = self.hospital_dates[1:]
self.hospital_times = self.hospital_times[1:]
self.log_likelihood = None
self.log = structlog.getLogger(Rt_Inference_Target=self.display_name)
self.log.info(event="Running:")
def __init__(self,
fips,
window_size=7,
kernel_std=2,
r_list=np.linspace(0, 10, 501),
process_sigma=0.15,
ref_date=datetime(year=2020, month=1, day=1),
confidence_intervals=(0.68, 0.75, 0.90)):
self.fips = fips
self.r_list = r_list
self.window_size = window_size
self.kernel_std = kernel_std
self.process_sigma = process_sigma
self.ref_date = ref_date
self.confidence_intervals = confidence_intervals
if len(fips) == 2: # State FIPS are 2 digits
self.agg_level = AggregationLevel.STATE
self.state_obj = us.states.lookup(self.fips)
self.state = self.state_obj.name
self.geo_metadata = load_data.load_county_metadata_by_state(self.state).loc[self.state].to_dict()
self.times, self.observed_new_cases, self.observed_new_deaths = \
load_data.load_new_case_data_by_state(self.state, self.ref_date)
self.hospital_times, self.hospitalizations, self.hospitalization_data_type = \
load_data.load_hospitalization_data_by_state(self.state_obj.abbr, t0=self.ref_date)
self.display_name = self.state
else:
self.agg_level = AggregationLevel.COUNTY
self.geo_metadata = load_data.load_county_metadata().set_index('fips').loc[fips].to_dict()
self.state = self.geo_metadata['state']
self.state_obj = us.states.lookup(self.state)
self.county = self.geo_metadata['county']
if self.county:
self.display_name = self.county + ', ' + self.state
else:
self.display_name = self.state
# TODO Swap for new data source.
self.times, self.observed_new_cases, self.observed_new_deaths = \
load_data.load_new_case_data_by_fips(self.fips, t0=self.ref_date)
self.hospital_times, self.hospitalizations, self.hospitalization_data_type = \
load_data.load_hospitalization_data(self.fips, t0=self.ref_date)
logging.info(f'Running Rt Inference for {self.display_name}')
self.case_dates = [ref_date + timedelta(days=int(t)) for t in self.times]
if self.hospitalization_data_type:
self.hospital_dates = [ref_date + timedelta(days=int(t)) for t in self.hospital_times]
self.default_parameters = ParameterEnsembleGenerator(
fips=self.fips,
N_samples=500,
t_list=np.linspace(0, 365, 366)
).get_average_seir_parameters()
# Serial period = Incubation + 0.5 * Infections
self.serial_period = 1 / self.default_parameters['sigma'] + 0.5 * 1 / self.default_parameters['delta']
# If we only receive current hospitalizations, we need to account for
# the outflow to reconstruct new admissions.
if self.hospitalization_data_type is load_data.HospitalizationDataType.CURRENT_HOSPITALIZATIONS:
los_general = self.default_parameters['hospitalization_length_of_stay_general']
los_icu = self.default_parameters['hospitalization_length_of_stay_icu']
hosp_rate_general = self.default_parameters['hospitalization_rate_general']
hosp_rate_icu = self.default_parameters['hospitalization_rate_icu']
icu_rate = hosp_rate_icu / hosp_rate_general
flow_out_of_hosp = self.hospitalizations[:-1] * ((1 - icu_rate) / los_general + icu_rate / los_icu)
# We are attempting to reconstruct the cumulative hospitalizations.
self.hospitalizations = np.diff(self.hospitalizations) + flow_out_of_hosp
self.hospital_dates = self.hospital_dates[1:]
self.hospital_times = self.hospital_times[1:]
self.log_likelihood = None
def __init__(
self,
fips,
ref_date=datetime(year=2020, month=1, day=1),
min_deaths=2,
n_years=1,
cases_to_deaths_err_factor=0.5,
hospital_to_deaths_err_factor=0.5,
percent_error_on_max_observation=0.5,
with_age_structure=False,
):
# Seed the random state. It is unclear whether this propagates to the
# Minuit optimizer.
np.random.seed(seed=42)
self.fips = fips
self.ref_date = ref_date
self.days_since_ref_date = (dt.date.today() - ref_date.date() -
timedelta(days=7)).days
# ndays end of 2nd ramp may extend past days_since_ref_date w/o penalty on chi2 score
self.days_allowed_beyond_ref = 0
self.min_deaths = min_deaths
self.t_list = np.linspace(0, int(365 * n_years),
int(365 * n_years) + 1)
self.cases_to_deaths_err_factor = cases_to_deaths_err_factor
self.hospital_to_deaths_err_factor = hospital_to_deaths_err_factor
self.percent_error_on_max_observation = percent_error_on_max_observation
self.t0_guess = 60
self.with_age_structure = with_age_structure
if len(fips) == 2: # State FIPS are 2 digits
self.agg_level = AggregationLevel.STATE
self.state_obj = us.states.lookup(self.fips)
self.state = self.state_obj.name
(
self.times,
self.observed_new_cases,
self.observed_new_deaths,
) = load_data.load_new_case_data_by_state(self.state,
self.ref_date)
(
self.hospital_times,
self.hospitalizations,
self.hospitalization_data_type,
) = load_data.load_hospitalization_data_by_state(
self.state_obj.abbr, t0=self.ref_date)
(
self.icu_times,
self.icu,
self.icu_data_type,
) = load_data.load_hospitalization_data_by_state(
self.state_obj.abbr,
t0=self.ref_date,
category=HospitalizationCategory.ICU)
self.display_name = self.state
else:
self.agg_level = AggregationLevel.COUNTY
geo_metadata = load_data.load_county_metadata().set_index(
"fips").loc[fips].to_dict()
state = geo_metadata["state"]
self.state_obj = us.states.lookup(state)
county = geo_metadata["county"]
if county:
self.display_name = county + ", " + state
else:
self.display_name = state
# TODO Swap for new data source.
(
self.times,
self.observed_new_cases,
self.observed_new_deaths,
) = load_data.load_new_case_data_by_fips(self.fips,
t0=self.ref_date)
(
self.hospital_times,
self.hospitalizations,
self.hospitalization_data_type,
) = load_data.load_hospitalization_data(self.fips,
t0=self.ref_date)
(
self.icu_times,
self.icu,
self.icu_data_type,
) = load_data.load_hospitalization_data(
self.fips,
t0=self.ref_date,
category=HospitalizationCategory.ICU)
self.cases_stdev, self.hosp_stdev, self.deaths_stdev = self.calculate_observation_errors(
)
self.set_inference_parameters()
self.model_fit_keys = [
"R0",
"eps",
"t_break",
"eps2",
"t_delta_phases",
"log10_I_initial",
]
self.SEIR_kwargs = self.get_average_seir_parameters()
self.fit_results = None
self.mle_model = None
self.chi2_deaths = None
self.chi2_cases = None
self.chi2_hosp = None
self.dof_deaths = None
self.dof_cases = None
self.dof_hosp = None
