Skip to main content

Analysis

'R'eady to go...​

To analyze the data ouput of the QExE tool, we've provided some R code for some data exploration.

Loading a Results File​

library(jsonlite)
library(tidyverse)

results_js <- fromJSON("subj_id_unknown_1809/results_id_unknown_1809.json", flatten = TRUE) # Load the JSON results file
df <- as.data.frame(results_js$ResultsFile$TestResults) # Select the JSON nest with the results

df_long <- df %>%
  pivot_longer(
    cols = starts_with("Rating.Con_"),  # Select columns that start with 'Rating.Con_'
    names_to = "Condition",             # Name of the new key column
    values_to = "Rating"                # Name of the new value column
  )

df_long <- df_long %>%
  mutate(Condition = str_replace(Condition, "Rating.Con_", "")) # Remove flattened JSON names

df_long <- df_long %>% # Revalue the column vectors to factors or numeric
  mutate(
    SceneID = as.factor(SceneID),
    UnitySceneID = as.factor(UnitySceneID),
    VideoFile = as.factor(VideoFile),
    VideoID = as.factor(VideoID),
    Condition = as.factor(Condition),
    Rating = as.numeric(Rating)
  )

df_long$SubjectID <- 1 # Add a subject ID

Loading a Questionnaire file​

library(jsonlite)
library(tidyverse)

questionnaire_js <- fromJSON("subj_id_unknown_1809/questionnaire_id_unknown_1809.json", flatten = TRUE)
questionnaire_df <- as.data.frame(questionnaire_js$QuestionnaireFile$QuestionnaireResults)

df_long <- questionnaire_df %>%
  pivot_longer(
    cols = starts_with("Rating."), 
    names_to = "NASATLX",             
    values_to = "Rating"            
  )

df_long <- df_long %>%
  mutate(NASATLX = str_replace(NASATLX, "Rating.", ""))

df_long <- df_long %>%
  mutate(
    SceneID = as.factor(SceneID),
    UnitySceneID = as.factor(UnitySceneID),
    VideoFile = as.factor(VideoFile),
    VideoID = as.factor(VideoID),
    NASATLX = as.factor(NASATLX),
    Rating = as.numeric(Rating)
  )

df_long$SubjectID <- 1

Load a Behavior Tracking File​

The behavior file contains a lot of data that you can parse into different dataframes for analysis. Here's an example loading the .txt file and parsing the user and object transform data into separate df's.

library(dplyr)
library()

df_this_subject_behavior = read.table("subj_id_unknown_1809/behaviour_id_unknown_1809.txt", header = FALSE, sep = " ", dec = ".", fill = TRUE)
df_this_subject_behavior <- dplyr::select(df_this_subject_behavior, V1, V2, V4, V5, V9, V12, V15, V18, V21, V24)
  
# df_this_subject_behavior$SubjID <- c(gsub(".txt","",subjectFilesInFolder[9]))

# Extract user-transform messages into new df
df_behavior_userPose <- dplyr::filter(df_this_subject_behavior, 
                                      V5 == "<OSCMessage:/stream/userpose/rot>" 
                                      | V5 == "<OSCMessage:/stream/userpose/pos>")

# Extract object-transform messages into new df
df_behavior_objects <- dplyr::filter(df_this_subject_behavior, 
                                     V5 == "<OSCMessage:/stream/objectspose/pos/>" 
                                     | V5 == "<OSCMessage:/stream/objectspose/rot/>" )

# Rename the column vectors. 
colnames(df_behavior_userPose) <- c("Scene", "TimeElapsed", "Video", "OSCMessage", "X", "Y", "Z")
colnames(df_behavior_objects) <- c("Scene", "TimeElapsed", "Video", "OSCMessage", "ObjectIndex", "X", "Y", "Z")

# User behavior
# Remove characters from numeric variable entries 
df_behavior_userPose$X <- gsub(",", ".", as.character(df_behavior_userPose$X))
df_behavior_userPose$X = substring(df_behavior_userPose$X,1, nchar(df_behavior_userPose$X)-1)
df_behavior_userPose$Y <- gsub(",", ".", as.character(df_behavior_userPose$Y))
df_behavior_userPose$Y = substring(df_behavior_userPose$Y,1, nchar(df_behavior_userPose$Y)-1)
df_behavior_userPose$Z <- gsub(",", ".", as.character(df_behavior_userPose$Z))
df_behavior_userPose$Z <- gsub(")", "", as.character(df_behavior_userPose$Z))
df_behavior_userPose$TimeElapsed <- gsub(",", ".", as.character(df_behavior_userPose$TimeElapsed))

# Convert vectors to factors and numeric values
df_behavior_userPose$Scene <- as.factor(df_behavior_userPose$Scene)
df_behavior_userPose$TimeElapsed <- as.numeric(df_behavior_userPose$TimeElapsed)
df_behavior_userPose$X <- as.numeric(df_behavior_userPose$X)
df_behavior_userPose$Y <- as.numeric(df_behavior_userPose$Y)
df_behavior_userPose$Z <- as.numeric(df_behavior_userPose$Z)

# Remove obsolete columns
df_behavior_userPose <- df_behavior_userPose[,c(1,2,3,4,5,6,7)]

# Objects
# Remove characters from numeric variable entries 
df_behavior_objects$X <- gsub(",", ".", as.character(df_behavior_objects$X))
df_behavior_objects$X = substring(df_behavior_objects$X,1, nchar(df_behavior_objects$X)-1)
df_behavior_objects$Y <- gsub(",", ".", as.character(df_behavior_objects$Y))
df_behavior_objects$Y = substring(df_behavior_objects$Y,1, nchar(df_behavior_objects$Y)-1)
df_behavior_objects$Z <- gsub(",", ".", as.character(df_behavior_objects$Z))
df_behavior_objects$Z = substring(df_behavior_objects$Z,1, nchar(df_behavior_objects$Z)-1)
df_behavior_objects$TimeElapsed <- gsub(",", ".", as.character(df_behavior_objects$TimeElapsed))
df_behavior_objects$ObjectIndex <- gsub(",", "", as.character(df_behavior_objects$ObjectIndex))

df_behavior_objects$Scene <- as.factor(df_behavior_objects$Scene)
df_behavior_objects$TimeElapsed <- as.numeric(df_behavior_objects$TimeElapsed)
df_behavior_objects$ObjectIndex <- as.numeric(df_behavior_objects$ObjectIndex)
df_behavior_objects$X <- as.numeric(df_behavior_objects$X)
df_behavior_objects$Y <- as.numeric(df_behavior_objects$Y)
df_behavior_objects$Z <- as.numeric(df_behavior_objects$Z)

# Remove obsolete columns
df_behavior_objects <- df_behavior_objects[,c(1,2,3,4,5,6,7,8)]

df_behavior_userPose$SubjectID <- 1
df_behavior_objects$SubjectID <- 1