prior/posterior predictive checks

[AugustHao]

add function to plot trajectories for reff, infection, and cases, with different underpinning prior/posterior values for parameters visible on plot (in title or label or something), for internal diagnostic checks

start with something like this

# prior simulation on reff

sims <- calculate(infections_mod,
                  reff,
                  nsim = n_sims)

# crop extreme values and initial noise
ylim <- c(0, 3)
xlim <- c(10, n_days)

plot(sims$reff[1, , ][,1] ~ days,
     ylim = ylim,
     xlim = xlim,
     type = "n",
     main = "reff prior")
for (i in seq_len(n_sims)) {
  for (j in seq_len(n_states)) {
    lines(sims$reff[i, , ][,j] ~ days)
  }
}
abline(h = 1, lty = 2)
# prior sim on infections
plot(sims$infections_mod[1, , ][,1] ~ days,
     type = "n",
     main = "infections prior")
for (i in seq_len(n_sims)) {
  for (j in seq_len(n_states)) {
    lines(sims$infections_mod[i, , ][,j] ~ days)
  }
}

[SeniorKate]

Output growth rate so we can explore posteriors