I am struggling to figure out how to convert the BORIS output into one of the state sequence analysis formats that I can analyze with TraMineR.
The BORIS outputs are basically tables that look like this:
File Time Behavior Status
1 K8121319_feed3_01 0.000 Approach START
2 K8121319_feed3_01 393.225 Approach STOP
3 K8121319_feed3_01 393.226 Out-of-Frame START
4 K8121319_feed3_01 426.003 Out-of-Frame STOP
5 K8121319_feed3_01 442.006 Approach START
6 K8121319_feed3_01 465.755 Approach STOP
7 K8121319_feed3_01 465.756 Avoid START
8 K8121319_feed3_01 513.255 Avoid STOP
9 K8121319_feed3_01 513.256 Explore START
10 K8121319_feed3_01 746.577 Explore STOP
It seems like it would be possible to convert to the SPELL sequence format using dplyr, but I can't figure out how. Has anyone used these two softwares together?
The SPELL format would look like this:
File Behavior Start Stop
1 K8121319_feed3_01 Approach 0.000 393.225
2 K8121319_feed3_01 OOF 393.226 426.003
3 K8121319_feed3_01 Approach 426.006 465.755
4 K8121319_feed3_01 Avoid 465.756 513.255
5 K8121319_feed3_01 Explore 513.256 746.577
I have been trying to use dplyr::spread to do this.
Edit: here is the result of dput(data1[1:20,])
structure(list(File = c("K8121319_feed3_01", "K8121319_feed3_01",
"K8121319_feed3_01", "K8121319_feed3_01", "K8121319_feed3_01",
"K8121319_feed3_01", "K8121319_feed3_01", "K8121319_feed3_01",
"K8121319_feed3_01", "K8121319_feed3_01", "K8121319_feed3_02",
"K8121319_feed3_02", "K8121319_feed3_02", "K8121319_feed3_02",
"K8121319_feed3_02", "K8121319_feed3_02", "K8121319_feed3_02",
"K8121319_feed3_02", "K8121319_feed3_02", "K8121319_feed3_02"
), Time = c(0, 393.225, 393.226, 426.003, 442.006, 465.755, 465.756,
513.255, 513.256, 746.577, 0, 29.85, 29.851, 66.6, 66.601, 292.646,
292.647, 362.208, 362.209, 442.456), Behavior = c("Approach",
"Approach", "Out-of-Frame", "Out-of-Frame", "Approach", "Approach",
"Avoid", "Avoid", "Explore", "Explore", "Approach", "Approach",
"Avoid", "Avoid", "Approach", "Approach", "Avoid", "Avoid", "Approach",
"Approach"), Status = c("START", "STOP", "START", "STOP", "START",
"STOP", "START", "STOP", "START", "STOP", "START", "STOP", "START",
"STOP", "START", "STOP", "START", "STOP", "START", "STOP")), row.names = c(NA,
20L), class = "data.frame")
Edit: dput for part of df with repeated states
dput(data1[360:370,])
structure(list(File = c("K8121819_feed3_13", "K8121819_feed3_13",
"K8121819_feed3_13", "K8121819_feed3_13", "K8121819_feed3_13",
"K8121819_feed3_14", "K8121819_feed3_14", "K8121819_feed3_14",
"K8121819_feed3_14", "K8121819_feed3_14", "K8121819_feed3_14"
), Time = c(700.311, 700.312, 720.311, 742.851, 754.339, 0, 32.124,
32.125, 47.14, 47.141, 84.671), Behavior = c("Approach", "Avoid",
"Avoid", "Avoid", "Avoid", "Avoid", "Avoid", "Explore", "Explore",
"Approach", "Approach"), Status = c("STOP", "START", "STOP",
"START", "STOP", "START", "STOP", "START", "STOP", "START", "STOP"
)), row.names = 360:370, class = "data.frame")
I question your statement that the SPELL format can be used with continuous data, because providing a double to seqdef results in an error that the beginning and end columns must be integer.
Hopefully this will get you started though:
Edit: Now to potentially fix duplicated Behavior states:
library(TraMineR)
library(tidyverse)
library(data.table)
data.long <- data1 %>%
mutate(id = rleid(Behavior),
Behavior = str_replace_all(Behavior,pattern = "-", replacement = "")) %>%
group_by(File,id) %>%
dplyr::filter(Time == min(Time) | Time == max(Time)) %>%
pivot_wider(id_cols = c("File","Behavior", "id"),
names_from = "Status",
values_from = "Time") %>%
mutate(START = 1L+as.integer(floor(START)),
STOP = 1L+as.integer(floor(STOP))) %>%
as.data.frame()
data.long
# File Behavior id START STOP
#1 K8121319_feed3_01 Approach 1 1 394
#2 K8121319_feed3_01 OutofFrame 2 394 427
#3 K8121319_feed3_01 Approach 3 443 466
#4 K8121319_feed3_01 Avoid 4 466 514
#5 K8121319_feed3_01 Explore 5 514 747
#6 K8121319_feed3_02 Approach 6 1 30
#7 K8121319_feed3_02 Avoid 7 30 67
#8 K8121319_feed3_02 Approach 8 67 293
#9 K8121319_feed3_02 Avoid 9 293 363
#10 K8121319_feed3_02 Approach 10 363 443
I removed the - because it was causing problems with seqstatl, and I added 1 because apparently the package authors thought 0 not allowed. I used rleid from the data.table package because it saved a lot of typing trying to use base R's rle.
Now we can use seqdef:
data.SPELL <- seqdef(data = data.long,
var = c("File", "START", "STOP", "Behavior"),
informat = "SPELL",
labels = seqstatl(data.long$Behavior),
states = seq_along(seqstatl(data.long$Behavior)),
process = FALSE)
data.SPELL
#K8121319_feed3_01 1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-4-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3-3
#K8121319_feed3_02 1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-2-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1-1