estimatepast_RKI_timeseries: estimatepast_RKI_timeseries

In nevrome/covid19germany: Load, Visualise and Analyse Daily Updated Data on the COVID-19 Outbreak in Germany

estimatepast_RKI_timeseries

Description

This function implements two sequential methods to estimate the real number of infected based on the current cumulative death count as presented by Tomas Pueyo in this (https://medium.com/@tomaspueyo/coronavirus-act-today-or-people-will-die-f4d3d9cd99ca) blog post.

Usage

estimatepast_RKI_timeseries(
  x,
  ...,
  prop_death,
  mean_days_until_death,
  doubling_time
)

Arguments

x	data.frame. RKI data as downloaded with get_RKI_timeseries
...	variable names. One or multiple grouping columns of x, so Bundesland, Landkreis, Gender or Age
prop_death	numeric. Probability of death
mean_days_until_death	integer. Mean number of days from infection to death (in case of death)
doubling_time	numeric. Mean number of days for the number of infected to double

Details

EstimationCumNumberIllPast, the actual number of infected, is calculated with the current cumulative number of deaths CumNumberDead, the death rate prop_death and the average number of days mean_days_until_death from infection to death (in case of death). This approach only allows to estimate values at least mean_days_until_death days in the past. EstimationCumNumberIllPresent employs the last value in EstimationCumNumberIllPast to estimate the number of actually infected people beyond the mean_days_until_death threshold with a simple exponential growth model considering doubling_time. With EstimationCumNumberIllPast, EstimationCumNumberIllPresent and prop_death we can calculate an expected number of deaths EstimationCumNumberDeadFuture.

Examples

rki_timeseries <- get_RKI_timeseries()

estimatepast_RKI_timeseries(
  rki_timeseries, 
  prop_death = 0.01, mean_days_until_death = 17, doubling_time = 4
)

estimatepast_RKI_timeseries(
  rki_timeseries, Bundesland, 
  prop_death = 0.03, mean_days_until_death = 17, doubling_time = 3
)

nevrome/covid19germany documentation built on Nov. 2, 2022, 12:25 a.m.