Extract Bayesian posterior predicted outputs for all subjects in an excel/csv

[ChrisKonig]

Hi all,
I am looking into a method to generate an excel or csv doc for the bayesian post predicted outputs for all subjects on an individual basis. Appreciate any help.

[mnneely]

Tidyverse can help you here. As an example...

library(tidyverse) #ensures you have necessary functions loaded

run1 <- PM_load(1) #assumes you have a completed run 1 in the Runs folder
run1$op$data %>% 
filter(pred.type == "post", icen == "median") %>% #filters to posterior predictions based on median of the Bayesian posterior
select(1:4) %>% #select the first 4 columns: id, time, obs, pred
write_csv(file = "myfile.csv") #write the object to a file

The above code assumes only one output equation. You'd have to amend your filter and select if you had more than one output. If you wanted the posterior predictions at frequent (default 5 per hour) intervals plus at the time of observations, use this code:

run1 <- PM_load(1) 
run1$post$data %>% 
filter(icen == "median") %>% 
select(id, time, pred) %>% 
write_csv(file = "myfile.csv") #write the object to a file

The filtering and selecting are slightly different since the $op and $post data objects are also slightly different.