Figure 7.7.

[ASKurz]

I can't seem to get the name to work dynamically (e.g., paste("fig", w, sep = "_") returns an error). I would like to code this such that the loop returns three objects fig_-1, fig_0 and fig_1, which I could save or put into the multiplot() function.

Here's my curent attempt:

# loop over values of waterC and plot predictions
shade.seq <- -1:1

for(w in -1:1){
  # defining the subset of the original data
  dt <- d[d$water.c == w, ]
  # defining our new data
  nd <- tibble(water.c = w, shade.c = shade.seq)
  # using our sampling skills, like before
  fit.7.9 <- fitted(b7.9, newdata = nd) %>%
    as_tibble() %>%
    bind_cols(nd)
  
  # specifying our custom plot
  fig <- ggplot() + 
    theme_pander() + 
    geom_ribbon(data = fit.7.9, 
                aes(x = shade.c,
                    ymin = `2.5%ile`,
                    ymax = `97.5%ile`), 
                fill = "#CC79A7", alpha = 1/5) +
    geom_line(data = fit.7.9, aes(x = shade.c, y = Estimate), 
              color = "#CC79A7") +
    geom_point(data = dt, aes(x = shade.c, y = blooms),
               color = "#CC79A7") +
    coord_cartesian(xlim = c(-1, 1), ylim = c(0, 350)) +
    scale_x_continuous(breaks = c(-1, 0, 1)) +
    labs(x = "Shade (centered)", y = "Blooms", 
         title = paste("Water (centered) =", w)) +
    theme(text = element_text(family = "Times"))
  
  # plotting that joint
  plot(fig)
}

[paul-buerkner]

Maybe the assign() function could be helpful to you. The name fig_-1 is invalid though.

[ASKurz]

Ah, that makes sense. Thanks Paul!