DiseasySeason

library(diseasy)
#> Loading required package: diseasystore
#> 
#> Attaching package: 'diseasy'
#> The following objects are masked from 'package:diseasystore':
#> 
#>     %.%, diseasyoption

Introduction

The DiseasySeason is the module responsible for the implementation of different models for the seasonal dependency of the disease.

Configuring the module

The module can be initialized without setting any parameters. Note that some models require the parameter reference_date to be set which can be done during initialization of the module.

s <- DiseasySeason$new(reference_date = as.Date("2020-03-01"))

To see the season models that comes with the module we can query the module.

s$available_season_models
#> [1] "constant_season" "cosine_season"   "covid_season_v1" "covid_season_v2"

By default, the module initializes with the constant_season() model. To check the current status of the module, including which model is set, the $describe() method can be used:

s$describe()
#> # Season model ###############################################
#> Season model: constant_season
#> Constant (no) seasonality model.
#> Risk of infection constant through year
#> Reference date: 2020-03-01

Once a season model is set, interfacing with the model happens through the $model_t() and $model_date() methods.

t <- 0:365
plot(t, purrr::map_dbl(t, s$model_t),
     type = "l", lwd = 2,
     xlab = "Days past reference date",
     ylab = "Relative effect of season",
     ylim = c(0, 1.25), xaxs = "i", yaxs = "i")

d <- s$reference_date + 0:365
plot(d, purrr::map_dbl(d, s$model_date),
     type = "l", lwd = 2,
     xlab = "",
     ylab = "Relative effect of season",
     ylim = c(0, 1.25), xaxs = "i", yaxs = "i")

To use the other models for season, the module provides functions of the form $use_*_season(). For example, to configure the module to use a cosine season instead we can use $use_cosine_season()

s$use_cosine_season()

Once re-configured the models stored in $model_t and $model_date change to the new model:

d <- s$reference_date + 0:365
plot(d, purrr::map_dbl(d, s$model_date),
     type = "l", lwd = 2,
     xlab = "",
     ylab = "Relative effect of season",
     ylim = c(0, 1.25), xaxs = "i", yaxs = "i")

Note that all models, other than constant_season has a scale parameter. The scale parameter sets the relative difference between minimum and maximum season effect in the model. For example, if scale = 0.25, the effect of season should never drop below 25% of the maximum.

s$set_scale(0.25)

d <- s$reference_date + 0:365
plot(d, purrr::map_dbl(d, s$model_date),
     type = "l", lwd = 2,
     xlab = "",
     ylab = "Relative effect of season",
     ylim = c(0, 1.25), xaxs = "i", yaxs = "i")

The models may have additional parameters that control their shape. See the documentation of the module to learn more about the specific models.