LOPUS Script

LOPUS.R

# The following code was written by Jake Kaupp at Queen's University, Kingston, Canada
# for looking at LOPUS observation data
# Date: February 2017

library(jsonlite)
library(plyr)
library(tidyverse)
library(lubridate)

# Data file is located in the same direoctory as this script
gorp_file <- list.files(pattern = "json") %>% 
  fromJSON(simplifyDataFrame = TRUE) 

# Change unix time to ms
observations <- gorp_file$observations %>% 
  mutate_each(funs(./1000), starts_with("timestamp"))

# Convert start and stop to POSIXct objects
observations$timestamp_on <- as.POSIXct(observations$timestamp_on, origin = "1970-01-01", tz = "EST")

observations$timestamp_off <- as.POSIXct(observations$timestamp_off, origin = "1970-01-01", tz = "EST")

# Find the minimum time
min_time <- min(observations$timestamp_on)
max_time <- max(observations$timestamp_off)

# Create summary data
clean_data <- observations %>% 
  mutate(span = as.period(interval(timestamp_on, timestamp_off), unit = "seconds")) %>%
  mutate_at(c("timestamp_off","timestamp_on"), function(x) difftime(x, min_time, units = "secs")) %>% 
  separate_rows(value, sep = ",")
  

axis_order <- c("Typical-L",
                "Typical-Lab",
                "Typical-TQ",
                "Interactive-SQ",
                "Interactive-1o1_SQ",
                "Interactive-WC",
                "Interactive-Prd",
                "Interactive-SP",
                "Interactive-SI",
                "Noninstructive-SL",
                "Noninstructive-W_students",
                "Noninstructive-Other_student")

summary_data <- clean_data %>% 
  group_by(key) %>%
  summarize(count = n(),
            duration = sum(span, na.rm = TRUE)) %>%
  filter(!grepl("Comment", key, ignore.case = TRUE)) %>%
  mutate(fill = case_when(grepl("Typical", .$key) ~ "blue",
                          grepl("Interactive", .$key) ~ "black",
                          grepl("Noninstructive", .$key) ~ "purple",
                          TRUE ~ "grey"))

# Write it to a csv
summary_data %>% 
  write.csv("LOPUS Data.csv")


# Bar plot of summary data by count
ggplot(summary_data, aes(x = key, y = count)) +
  geom_col(aes(fill = fill, color = fill)) +
  scale_x_discrete(limits = rev(axis_order), breaks = rev(axis_order), drop = FALSE) +
  scale_color_identity() +
  scale_fill_identity() +
  coord_flip()

# Ordered bar plot of actitivites in summary data
ggplot(summary_data, aes(x = reorder(key, count), y = count)) +
  geom_col() +
  coord_flip()


#Gorp Time span 

max_time <- as.period(interval(min_time, max_time), unit = "minutes") %>% 
  as.numeric()/60

max <- round_any(max_time, 10, ceiling)

ggplot(clean_data, aes(y = key, group = key)) +
  geom_segment(aes(x = timestamp_on, xend = timestamp_off, y = factor(key), yend = factor(key)), size = 5) +
  scale_x_continuous(breaks = seq(0,max,2)*60, labels = seq(0,max,2)) +
  theme_minimal() +
  theme(panel.grid.minor = element_blank())