Script used to calculate attendance to classes for students over Zoom.

zoom_attendance.R

require(tidyverse)
require(lubridate)
setwd("") #

files <- (Sys.glob("*.csv"))

topics <- lapply(files, read.csv, nrows=2, header=T, encoding="UTF-8")
topics <- lapply(topics, function(x) x["Topic"][1,]) %>% unlist()

ldf <- lapply(files, read.csv, skip=3, header=T, encoding="UTF-8")
the_colnames <- c("name", "email", "join_time", "leave_time", "duration", "guest", "topic")
for (i in 1:length(ldf)){
  ldf[[i]]["topic"] <- topics[i]
  colnames(ldf[[i]]) <- the_colnames
}
l2 <- map(ldf,~data.frame(.))
l3 <-map_dfr(l2,~mutate_all(.,as.character))  
l4<-as_tibble(l3)

l4 <-  l4 %>% select(name, join_time, leave_time)
l4$join_time <- l4$join_time %>% mdy_hms(tz = "America/El_Salvador")
l4$leave_time <- l4$leave_time %>% mdy_hms(tz = "America/El_Salvador")

l4$hms_join_time = hms::as_hms(l4$join_time)
l4$hms_leave_time = hms::as_hms(l4$leave_time)
l4$wday = wday(l4$join_time, week_start = 1)

l4 <- l4 %>% mutate(name = str_replace_all(name, "[^a-zA-Z\u00C0-\u017F \\n]", ""))
l4$name <- l4$name %>% str_trim()

report <- function(l4, class) {
  entered_late = vector()
  left_early = vector()
  duration = vector()
  
  l5 <- l4 %>% filter(wday == class[1] & hms(hms_leave_time) < hms(class[3]) + minutes(60) & hms(hms_join_time) > hms(class[2]) - minutes(90))
  
  l5$date <- date(l5$join_time)
  
  l5 <- l5 %>% unique() %>%  mutate(duration = round(interval(hms_join_time, hms_leave_time, tzone = "America/El_Salvador")/minutes(1), 0))

  l5 <- l5 %>% 
    group_by(name, date) %>% 
    summarize(total_duration = sum(duration), across()) %>% 
    select(name, date, wday, hms_join_time, hms_leave_time, total_duration) %>% 
    unique()
  
  l6 <- l5 %>% mutate(on_time = hms(hms_join_time) < hms(class[2]) + minutes(30)) %>% 
    unique()

  l6 <- l6 %>% mutate(
    above_quantile = total_duration >= l5$total_duration %>% quantile(as.numeric(class[4])))

  l7 <- l6 %>%
    select(name, 
           date, 
           total_duration, 
           join_time=hms_join_time, 
           leave_time=hms_leave_time, 
           on_time, 
           above_quantile)

  l7 <- l7 %>%
    group_by(name) %>%
    # filter(attendances >= 3) %>% 
    unique()
  
  l7 <- left_join(l7, 
                  l7 %>% 
                    group_by(name, date, above_quantile) %>% 
                    summarize(on_time = sum(on_time) > 0) %>% 
                    mutate(real_attendance = on_time * above_quantile)
                  )
  
  return (l7)
}

# class_1 <- c(
#   "1",                  # día de la semana
#   "10:25:00",           # hora de inicio
#   "12:05:00",           # hora de fin
#   ".25"                 # percentil aceptado como asistencia
#   )
# class_2 <- c(
#   "5",                  # día de la semana
#   "10:25:00",           # hora de inicio
#   "12:05:00",           # hora de fin
#   ".25"                 # percentil aceptado como asistencia
# )

class_3 <- c(
  "2",                  # día de la semana
  "14:40:00",           # hora de inicio
  "16:20:00",           # hora de fin
  ".25"                 # percentil aceptado como asistencia
)
class_4 <- c(
  "3",                  # día de la semana
  "18:20:00",           # hora de inicio
  "20:00:00",           # hora de fin
  ".25"                 # percentil aceptado como asistencia
)

asistencias_1 <- tibble()
asistencias_2 <- tibble()
asistencias_1 <- report(l4, class_3) 
asistencias_2 <- report(l4, class_4) 

asistencias <- tibble()
asistencias <- bind_rows(asistencias_1, asistencias_2)

asistencias <- drop_na(asistencias_1) %>% filter(real_attendance == 1)

asistencias <- asistencias %>% 
  filter(on_time == T) %>% 
  group_by(name) %>%
  summarize(attendances = sum(attendances), across()) %>%
  distinct(name, date, attendances, total_duration)
 
# asistencias %>% ggplot(aes(date, name, fill=total_duration)) + 
#   geom_tile(aes(width=1.5, height=1.9), show.legend = FALSE) + 
#   scale_fill_gradientn(colours=c("white","green")) +
#   theme_minimal() +
#   labs(x="Fecha", y="Nombre") 

asistencias %>% ggplot(aes(date, name, fill=total_duration)) + 
  geom_point(color="#482374", show.legend = FALSE) + 
  theme_minimal() +
  labs(x="Fecha", y="Nombre") 

asistencias %>% group_by(name) %>% summarize(total = sum(real_attendance)) %>% arrange(desc(total)) %>%  View()