Vectorized Plots

Kyle Baron

2020-03-11

Set up

library(pmplots)
library(dplyr)

data <- pmplots_data_obs()

Vectorized plots

Some pmplots functions have been vectorized so that you can pass in a vector of inputs and get a list of plots back. This is a convenience feature that allows you to quickly make several plots. The intended use is to arrange those plots in a single page or multiple pages for display.

For example, we can plot WRES versus WT

wres_cont(data, x = "WT")

If we wanted to vectorize this plot and look at WRES versus WT, ALB, CRCL, and AST we would write

wres_cont(data, x = c("WT", "ALB", "CRCL", "AST"))

## [[1]]

## 
## [[2]]

## 
## [[3]]

## 
## [[4]]

And we get a list of plots back. This list can be arranged on the fly with

covs <- c(
  "WT//Weight (kg)", "ALB//Albumin", 
  "CRCL//Creatinine clearance", "AST//Aspartate aminotransferase"
)

wres_cont(data, x = covs) %>% pm_grid()

## Loading required namespace: cowplot

In this example, we also made the vector of inputs full col_label specification.

Another example are the eta_cont plots

id <- pmplots_data_id()


eta_cont(data, x = covs, y = "ETA1//ETA-CL") %>% pm_grid()

What plots are vectorized?

Basically any plot where the user is required to identify the column for plotting. So in the example, wres_cont asks the user to specify what the continuous variable is for the x-axis. This plot is vectorized. Same with wres_cat

wres_cat(data, x = c("STUDYc", "CPc")) %>% pm_grid()

In contrast, dv_pred is hard-wired to look for DV and PRED. It is designed to look for only one thing on each axis. This plot is not vectorized.

You can use pm_scatter_list as a vectorized function to vectorize anyting. So this allows us to do

pm_scatter_list(data, y  = "WRES", x = c("PRED", "IPRED", "TIME")) %>% pm_grid()

Or just use lapply or purrr::map

lapply(c("PRED", "IPRED", "TIME"), wres_cont, df = data, y = "WRES") %>% pm_grid()

Please see the help topic for each function to know if that function is vectorized or not.