Counting Peaks in R per group(s) of a data.frame

Question

This is a followup question to Counting peaks in r per group.

Reproducible data:

set.seed(949494)
Happiness <- round(runif(100, -100, 100))
ID <- rep(c("ID1", "ID2", "ID3", "ID4", "ID5"), 20)
Stimuli <- rep(1:4, 1)
DF <- data.frame(ID, Stimuli, Happiness)

Function to calculate the Happiness threshold per ID by using each ID's unique sd():

# 1SD
f.SD1 <- function(y) {
  SD1_thresh <- mean(y) + (1*sd(y))
  return(SD1_thresh)
}

Function to identify when Happiness is above (TRUE = 1) or below (FALSE = 0) the threshold:

# SD1
f.Peaks_SD1 <- function(X, thresh) {
  H_peaks_1 <- ifelse(X >= thresh ,1,0)
  return(H_peaks_1)
}

Now I want to group by ID and Stimuli so that I can determine average peaks per stimuli:

H_peaks_1_df <- DF %>% group_by(Stimuli, ID) %>% summarise(thresh_SD1 = f.SD1(Happiness), ttime = sum(Happiness > thresh_SD1), nP_H_SD1 = sum(diff(c(f.Peaks_SD1(Happiness, thresh = thresh_SD1), 0)) < 0))
H_peaks_1_df
summary(H_peaks_1_df)

Output:

The problem with this output is that the thresholds for the same ID are different because the sd() was calculated per ID per Stimuli. I want to calculate the sd() across all Stimuli per ID and then count peaks per Stimuli. So, the output H_peaks_1_df here is perfect (group_by(Stimuli, ID)), just the column "thresh_SD1" should be the same value for ID1, namely "58.5" which is correctly calculated when grouping only by ID.

Is it possible in dplyr to execute the "thresh_SD1" calculation via group_by(ID) and then count peaks and total time via group_by(Stimuli, ID) in simple code?

Thanks in advance!

Answer 1

Yes, it is possible, Using head, mean or other to retrieve only one element from thresh_SD1.

H_peaks_1_df <- DF %>% group_by(ID) %>% 
  mutate(thresh_SD1 = f.SD1(Happiness)) %>% 
  group_by(ID, Stimuli) %>% 
    summarise(thresh_SD1 = head(thresh_SD1,1), ttime = sum(Happiness > thresh_SD1), nP_H_SD1 = sum(diff(c(f.Peaks_SD1(Happiness, thresh = thresh_SD1), 0)) < 0))
H_peaks_1_df


ID    Stimuli thresh_SD1 ttime nP_H_SD1
   <chr>   <int>      <dbl> <int>    <int>
 1 ID1         1       58.5     0        0
 2 ID1         2       58.5     2        2
 3 ID1         3       58.5     0        0
 4 ID1         4       58.5     1        1
 5 ID2         1       71.3     1        1
 6 ID2         2       71.3     1        1
 7 ID2         3       71.3     3        1
 8 ID2         4       71.3     0        0