Package index • epinowcast

Nowcast

Functions for nowcasting.

epinowcast(): Nowcast using partially observed data
plot(<epinowcast>): Plot method for epinowcast
summary(<epinowcast>): Summary method for epinowcast

Data converters

Functions for converting between data formats

enw_add_cumulative(): Calculate cumulative reported cases from incidence of new reports
enw_add_incidence(): Calculate incidence of new reports from cumulative reports
enw_aggregate_cumulative(): Aggregate observations over a given timestep for both report and reference dates.
enw_incidence_to_linelist(): Convert Aggregate Counts (Incidence) to a Line List
enw_linelist_to_incidence(): Convert a Line List to Aggregate Counts (Incidence)

Preprocess

Functions for preprocessing observations

enw_add_delay(): Add a delay variable to the observations
enw_add_max_reported(): Add the maximum number of reported cases for each reference_date
enw_add_metaobs_features(): Add common metadata variables
enw_assign_group(): Assign a group to each row of a data.table
enw_complete_dates(): Complete missing reference and report dates
enw_construct_data(): Construct preprocessed data
enw_extend_date(): Extend a time series with additional dates
enw_filter_delay(): Filter observations to have a consistent maximum delay period
enw_filter_reference_dates(): Filter by reference dates
enw_filter_report_dates(): Filter by report dates
enw_flag_observed_observations(): Flag observed observations
enw_impute_na_observations(): Impute NA observations
enw_latest_data(): Filter observations to the latest available reported
enw_metadata(): Extract metadata from raw data
enw_metadata_delay(): Calculate reporting delay metadata for a given maximum delay
enw_missing_reference(): Extract reports with missing reference dates
enw_preprocess_data(): Preprocess observations
enw_reporting_triangle(): Construct the reporting triangle
enw_reporting_triangle_to_long(): Recast the reporting triangle from wide to long format

Model modules

Modular model components

enw_expectation(): Expectation model module
enw_fit_opts(): Format model fitting options for use with stan
enw_missing(): Missing reference data model module
enw_obs(): Setup observation model and data
enw_reference(): Reference date logit hazard reporting model module
enw_report(): Report date logit hazard reporting model module

Helpers for defining model modules

Functions that help with the setup of model modules and that may be specific to individual modules

add_max_observed_delay(): Add maximum observed delay
add_pmfs(): Add probability mass functions
convolution_matrix(): Construct a convolution matrix
enw_dayofweek_structural_reporting(): Create day-of-week structural reporting pattern
enw_reference_by_report(): Construct a lookup of references dates by report
enw_reps_with_complete_refs(): Identify report dates with complete (i.e up to the maximum delay) reference dates
enw_structural_reporting_metadata(): Create structural reporting metadata grid
extract_obs_metadata(): Extract observation metadata
extract_sparse_matrix(): Extract sparse matrix elements
latest_obs_as_matrix(): Convert latest observed data to a matrix

Model tools

Generic functions that assist with combining and using modular models

enw_formula_as_data_list(): Format formula data for use with stan
enw_get_cache(): Retrieve Stan cache location
enw_model(): Load and compile the nowcasting model
enw_pathfinder(): Fit a CmdStan model using the pathfinder algorithm
enw_priors_as_data_list(): Convert prior data.frame to list
enw_replace_priors(): Replace default priors with user specified priors
enw_sample(): Fit a CmdStan model using NUTS
enw_set_cache(): Set caching location for Stan models
enw_stan_to_r(): Expose epinowcast stan functions in R
enw_unset_cache(): Unset Stan cache location
remove_profiling(): Remove profiling statements from a character vector representing stan code
write_stan_files_no_profile(): Write copies of the .stan files of a Stan model and its #include files with all profiling statements removed.

Postprocess

Functions for postprocessing model output

build_ord_obs(): Build the ord_obs data.table.
enw_add_latest_obs_to_nowcast(): Add latest observations to nowcast output
enw_nowcast_samples(): Extract posterior samples for the nowcast prediction
enw_nowcast_summary(): Summarise the posterior nowcast prediction
enw_posterior(): Summarise the posterior
enw_pp_summary(): Posterior predictive summary
enw_quantiles_to_long(): Convert summarised quantiles from wide to long format
enw_summarise_samples(): Summarise posterior samples
subset_obs(): Subset observations data table for either modelled dates or not-modelled earlier dates.

Plot

Functions for plotting postprocessed nowcast model output

enw_plot_nowcast_quantiles(): Plot nowcast quantiles
enw_plot_obs(): Generic quantile plot
enw_plot_pp_quantiles(): Plot posterior prediction quantiles
enw_plot_quantiles(): Generic quantile plot
enw_plot_theme(): Package plot theme
plot(<epinowcast>): Plot method for epinowcast

Model validation

Functions for validating model fits

as_forecast_sample(<epinowcast>): Convert an epinowcast object to a forecast_sample object

Formula design tools

Functions that assist with interpreting model formulas

as_string_formula(): Converts formulas to strings
construct_re(): Constructs random effect terms
construct_rw(): Constructs random walk terms
enw_formula(): Define a model using a formula interface
enw_manual_formula(): Define a model manually using fixed and random effects
parse_formula(): Parse a formula into components
re(): Defines random effect terms using the lme4 syntax
remove_rw_terms(): Remove random walk terms from a formula object
rw(): Adds random walks with Gaussian steps to the model.
rw_terms(): Finds random walk terms in a formula object
split_formula_to_terms(): Split formula into individual terms

Model design tools

Functions that assist with designing models

enw_add_cumulative_membership(): Add a cumulative membership effect to a data.frame
enw_add_pooling_effect(): Add a pooling effect to model design metadata
enw_design(): A helper function to construct a design matrix from a formula
enw_effects_metadata(): Extracts metadata from a design matrix
enw_one_hot_encode_feature(): One-hot encode a variable and column-bind it to the original data.table

Datasets

Package datasets used in examples and by users to explore the package functionality

enw_example(): Load a package example
germany_covid19_hosp: Hospitalisations in Germany by date of report and reference

Simulate Datasets

Tools for simulating datasets

enw_simulate_missing_reference(): Simulate observations with a missing reference date.

Check inputs

Functions to check the structure of user inputs

check_design_matrix_sparsity(): Check design matrix sparsity
check_group(): Check observations for reserved grouping variables
check_group_date_unique(): Check observations for uniqueness of grouping variables with respect to reference_date and report_date
check_max_delay(): Check appropriateness of maximum delay
check_module(): Check a model module contains the required components
check_modules_compatible(): Check that model modules have compatible specifications
check_numeric_timestep(): Check Numeric Timestep
check_observation_indicator(): Check observation indicator
check_quantiles(): Check required quantiles are present
check_timestep(): Check timestep
check_timestep_by_date(): Check timestep by date
check_timestep_by_group(): Check timestep by group

Utilities

Utility functions

coerce_date(): Coerce Dates
coerce_dt(): Coerce data.tables
date_to_numeric_modulus(): Convert date column to numeric and calculate its modulus with given timestep.
enw_rolling_sum(): Perform rolling sum aggregation
get_internal_timestep(): Get internal timestep
is.Date(): Check an object is a Date
stan_fns_as_string(): Read in a stan function file as a character string