epidplot: Visualize an epdemic outbreak

In eliaskrainski/emisc: Functions for personal utility

Visualize an epdemic outbreak

Description

visualize the cummulated time series and its derivative (news). Add smoothed lines of the news. Additional plots shows the 1st and 2nd derivatives of the smoothed, interpreted as the absolute increase and velocity.

Usage

epidplot(
  x,
  y,
  log10 = TRUE,
  which = list(1:2, c(3, 5), 4, 6),
  w = exp(-sqrt(1:14)) * 0.686,
  leg.args = NULL,
  lxlab = NULL,
  lylab = NULL,
  ask = prod(par("mfcol")) < length(which) && dev.interactive(),
  ...
)

Arguments

x	a date or numeric vector
y	the cummulated number of cases
log10	length one or two logical to indicate if accumulated and new series are shown in the log10 transformation.
which	integer vector or list of integers to specify which plots will be produced in each figure. See detais.
w	numeric vector to compute the expected number of new cases given past cases. See details.
leg.args	a named list with arguments passed to the legend of the first plot.
lxlab	a list with the xlab for each plot.
lylab	a list with the xlab for each plot.
ask	logical if ‘TRUE’, the user is _ask_ed before each plot, see par(ask=.).
...	additional arguments passed to hist

Details

This function produces up to 6 plots. 1st: accumulated, provided y, with smoothed line 2nd: new cases with smoothed line 3nd: derivative, change on the trend, of the smoothed line for new cases. 4th: derivative divided by the smoothed new cases line 5th: second derivative of the smoothed line for the new cases series, representing the velocity of the absolute change. 6th: effective reproduction number. The first and second plots can be joined into one. The smoothed line for the second plot is computed using the mgcv::gam function with mgcv::s() on x considering family=poisson(). The line for the 1st plot is the cumsum of the line being fitted for the 2nd plot. The last plot consider R_t = s_t/E_t where E_t = sum_j w_j s_t-j, for j = 1, ..., k=length(w). In this plot it will be shown the last (n-k) values for R_t, if type!='n'.

Value

up to six plots

Examples

## COVID19 cases in Nova Scotia, Canada
ns.cases <- c(5, 7, 12, 14, 15, 21, 28, 41,
 51, 68, 73, 90, 110, 122, 127, 147, 173,
 193, 207, 236, 262, 293, 310, 310, 342,
 407, 428, 445, 474, 517, 549, 579, 606,
 649, 675, 721, 737, 772, 827, 850, 865,
 873, 900, 915, 935, 947, 959, 963, 971,
 985, 991, 998, 1007, 1008, 1011, 1018,
 1019, 1020, 1024, 1026, 1034, 1037, 1040,
 1043, 1044, 1045, 1046, 1048, 1049, 1050,
 1051, 1052, 1053, 1055, 1055, 1056, 1056,
 1057, 1057, 1058, 1058, 1058, 1058, 1059,
 1059, 1060, 1061, 1061, 1061, 1061)
date <- as.Date('2020-03-15') + 1:length(ns.cases)
pw <- pgamma(0:14, shape=(5/3)^2, scale=3^2/5)
w <- diff(pw)/sum(diff(pw))
par(mfrow=c(2,2), mar=c(2,3,1,1),
    mgp=c(2, 0.5, 0), las=1)
epidplot(date, ns.cases, logbase=10, w=w)

eliaskrainski/emisc documentation built on Nov. 18, 2024, 11:02 a.m.