Worked on preprocessing; added trace variable; tried out modeling; prepared questions for PM

This commit is contained in:
Nora Wickelmaier 2023-08-14 16:57:03 +02:00
parent 9e3783cf1f
commit c63c0a8206
7 changed files with 683 additions and 49 deletions

103
README.md Normal file
View File

@ -0,0 +1,103 @@
# Offene Fragen
## Datenverständnis
* Welche Einheit haben x und y? Pixel?
* Welche Einheit hat scale?
* rotation wirklich degree?
* Nach welchem Zeitintervall resettet sich der Tisch wieder in die
Ausgangskonfiguration?
## Tisch-Software
* Gibt es Doku für die Bilder, die über die xml files hinausgeht? Sowas wie
ein Manual oder ähnliches?
* Gibt es evtl. irgendwo noch ein Tablet mit der Anwendung drauf?
* Was bedeuten die Farben der Topic Cards?
## Event Logs
* Wie gehen wir mit "nicht geschlossenen" Events um? Einfach rauslöschen?
- für Transform tendiere ich zu ja, weil sonst total uninteressant
- bei flipCard bin ich nicht so sicher... Aber man kann dann keine
duration berechnen, wäre NA
* Moves/scales/rotations ohne Veränderung würde ich auf jeden Fall
rauslöschen
* Es ist nicht möglich (bzw. ich weiß nicht wie) zusammengehörige Events
eineindeutig zu identifizieren
- nach Heuristik vorgehen? Doppelte Transformation start und stop einfach
raus?
- Daten sind nicht "fehlerfrei"; es gibt z.B. Transformation-Events wo
das Ende nicht geloggt wurde
* Wie identifiziere ich eine "Interaktionseinheit"?
- Was ist ein "case"?
- Eher grob über Zeitintervalle?
- Noch irgendeine andere Idee?
* Herausfinden, ob mehr als eine Person am Tisch steht?
- Sliding window, in der Anzahl von Artworks gezählt wird? Oder wie weit
angefasste Artworks voneinander entfernt sind?
- Man kann sowas schon "sehen" in den Logs - aber wie kann ich es
automatisiert rausziehen? Was ist meine Definition von
"Interaktionsboost"?
- Egal wie wir es machen, geht es auf den "Event-Log-Daten"?
* Anreicherung der Log-Daten mit weiteren Metadaten? Was wäre interessant?
- Metadata on artworks like, name, artist, type of artwork, epoch, etc.
ˆ - School vacations and holidays
ˆ - Special exhibits at the museum
ˆ - Number of visitors per day
ˆ - Age structure of visitors per day?
- ... ????
## HAUM
* Bei Sven noch mal nachhaken wegen Besucherzahlen?
# Problems and how I handled them
This lists some problems with the log data that required decisions. These
decisions influence the outcome and maybe even the data quality. Hence, I
tried to document how I handled these problems and explain the decisions I
made.
## Weird behavior of `time_ms` and neg. `duration`values
## Events that only close (`date.start` is NA)
## Timestamps repeat
## Popups from glossar cannot be assigned to a specific artwork
## Assign a case variable based on "time heuristic"
## A `move`event does not record any change
## Add moves to `trace` variable
# Reading list
* @Arizmendi2022 [$-$]
* @Bannert2014 [x]
* @Bousbia2010 [$-$]
* @Cerezo2020
* @GerjetsSchwan2021 [x]
* @Goldhammer2020
* @Guenther2007
* @HuberBannert2023 [x]
* @Kroehne2018
* @SchwanGerjets2021 [x]
* @vanderAalst2016 [Chap. 2, x]
* @vanderAalst2016 [Chap. 3]
* @vanderAalst2016 [Chap. 5, x]
* @Wang2019

View File

@ -4,8 +4,7 @@
#' date: "`r Sys.Date()`" #' date: "`r Sys.Date()`"
#' output: #' output:
#' html_document: #' html_document:
#' toc: true #' default
#' toc_float: true
#' pdf_document: #' pdf_document:
#' toc: true #' toc: true
#' number_sections: true #' number_sections: true
@ -17,8 +16,6 @@
#+ setup, include = FALSE #+ setup, include = FALSE
knitr::opts_chunk$set(warning = FALSE, message = FALSE) knitr::opts_chunk$set(warning = FALSE, message = FALSE)
#' # Preprocessing raw log files into data frame
#' The following events can be extracted from the log files: #' The following events can be extracted from the log files:
#' #'
#' ``` #' ```
@ -36,8 +33,8 @@ knitr::opts_chunk$set(warning = FALSE, message = FALSE)
#' Choose which folders with raw log files should be included: #' Choose which folders with raw log files should be included:
folders <- "all" #folders <- "all"
#folders <- "_2016b" folders <- "_2016b"
dirpaths <- paste0("../data/haum_logs_2016-2023/", folders) dirpaths <- paste0("../data/haum_logs_2016-2023/", folders)
@ -71,7 +68,7 @@ dat <- subset(dat, dat$logs != "")
d2 <- dim(dat)[1] d2 <- dim(dat)[1]
#' The files contain `r d1-d2` corrupt lines that were remooved from the data. #' The files contain `r d1-d2` corrupt lines that were remooved from the data.
#'
#' ### Extract relevant infos #' ### Extract relevant infos
@ -136,6 +133,8 @@ dat <- dat[order(dat$date), ]
## TODO: Replace artwork and popup numbers with informative strings ## TODO: Replace artwork and popup numbers with informative strings
write.table(dat, "../data/rawdata_logfiles.csv", #' ### Save data frame
write.table(dat, "../data/rawdata_logfiles_small.csv",
sep = ";", quote = FALSE, row.names = FALSE) sep = ";", quote = FALSE, row.names = FALSE)

View File

@ -222,3 +222,18 @@ lattice::barchart(counts, auto.key = TRUE)
#' can happen that the wrong tags have been put together (e.g., Transform #' can happen that the wrong tags have been put together (e.g., Transform
#' start and Transform stop); therefore, durations etc. are only heuristic #' start and Transform stop); therefore, durations etc. are only heuristic
#' ## Plots
counts <- table(as.Date(dat$date), dat$event)
lattice::barchart(counts, auto.key = TRUE)
start_events <- c("Transform start", "Show Info", "ShowPopup", "Artwork/OpenCard")
counts <- table(as.Date(dat$date[dat$event %in% start_events]),
dat$event[dat$event %in% start_events])
lattice::barchart(counts, auto.key = TRUE)

View File

@ -109,11 +109,16 @@ trans_wide$rotationDegree <- trans_wide$rotation.stop -
trans_wide$rotation.start trans_wide$rotation.start
trans_wide$scaleSize <- trans_wide$scale.stop - trans_wide$scale.start trans_wide$scaleSize <- trans_wide$scale.stop - trans_wide$scale.start
dat_trans <- dat_trans[trans_wide$distance != 0 & trans_wide$trace <- NA
trans_wide$card <- NA
trans_wide$popup <- NA
dat_trans <- trans_wide[trans_wide$distance != 0 &
trans_wide$rotationDegree != 0 & trans_wide$rotationDegree != 0 &
trans_wide$scaleSize != 0, trans_wide$scaleSize != 0,
c("event", "artwork", "date.start", "date.stop", c("event", "artwork", "trace", "date.start", "date.stop",
"time_ms.start", "time_ms.stop", "duration", "time_ms.start", "time_ms.stop", "duration",
"card", "popup",
"x.start", "y.start", "x.stop", "y.stop", "x.start", "y.start", "x.stop", "y.stop",
"distance", "scale.start", "scale.stop", "distance", "scale.start", "scale.stop",
"scaleSize", "rotation.start", "rotation.stop", "scaleSize", "rotation.start", "rotation.stop",
@ -128,7 +133,8 @@ summary(dat_trans)
# --> Hat er eine Erklärung dafür? # --> Hat er eine Erklärung dafür?
#plot(time_ms.stop ~ time_ms.start, dat_trans, type = "b") #plot(time_ms.stop ~ time_ms.start, dat_trans, type = "b")
plot(time_ms.stop ~ time_ms.start, dat_trans, col = rgb(red = 0, green = 0, blue = 0, alpha = 0.2)) plot(time_ms.stop ~ time_ms.start, dat_trans,
col = rgb(red = 0, green = 0, blue = 0, alpha = 0.2))
plot(date.stop ~ date.start, dat_trans[1:1000,], type = "b") plot(date.stop ~ date.start, dat_trans[1:1000,], type = "b")
@ -248,7 +254,7 @@ lut[sapply(lut$artwork, length) == 1, "glossar_file"]
# TODO: Fill in the ones that are associated with one artwork # TODO: Fill in the ones that are associated with one artwork
# --> Can't come up with something -- maybe ask AK??? # --> Can't come up with something -- maybe ask AK???
# TODO: How to check if one of the former "Show Infos" is correct on # TODO: How to check if one of the former "Show Infos" is correct
# --> Can't come up with something -- maybe ask AK??? # --> Can't come up with something -- maybe ask AK???
# for (file in lut$glossar_file) { # for (file in lut$glossar_file) {
@ -285,6 +291,8 @@ lut[sapply(lut$artwork, length) == 1, "glossar_file"]
# correct: 17940 # correct: 17940
# incorrect: 17963 # incorrect: 17963
# TODO: "glossar" entry should be changed to the corresponding artwork
# TODO: Add additional variable `glossar` with 0/1 or similar instead
# TODO: For now: Exclude not matched glossar entries # TODO: For now: Exclude not matched glossar entries
@ -314,10 +322,28 @@ flipCard_wide$duration <- flipCard_wide$time_ms.stop -
flipCard_wide$duration <- ifelse(flipCard_wide$duration < 0, flipCard_wide$duration <- ifelse(flipCard_wide$duration < 0,
NA, flipCard_wide$duration) NA, flipCard_wide$duration)
flipCard_wide$card <- NA
flipCard_wide$popup <- NA
flipCard_wide$x.start <- NA
flipCard_wide$x.stop <- NA
flipCard_wide$y.start <- NA
flipCard_wide$y.stop <- NA
flipCard_wide$distance <- NA
flipCard_wide$scale.start <- NA
flipCard_wide$scale.stop <- NA
flipCard_wide$scaleSize <- NA
flipCard_wide$rotation.start <- NA
flipCard_wide$rotation.stop <- NA
flipCard_wide$rotationDegree <- NA
dat_flipCard <- flipCard_wide[, c("event", "artwork", "trace", dat_flipCard <- flipCard_wide[, c("event", "artwork", "trace",
"date.start", "date.stop", "date.start", "date.stop",
"time_ms.start", "time_ms.stop", "time_ms.start", "time_ms.stop",
"duration")] "duration", "card", "popup",
"x.start", "y.start", "x.stop", "y.stop",
"distance", "scale.start", "scale.stop",
"scaleSize", "rotation.start",
"rotation.stop", "rotationDegree")]
rm(tmp, flipCard_wide) rm(tmp, flipCard_wide)
@ -344,10 +370,27 @@ openTopic_wide$duration <- ifelse(openTopic_wide$duration < 0,
# TODO: How to handle duration < 0 # TODO: How to handle duration < 0
# --> Replace with NA for now... # --> Replace with NA for now...
dat_openTopic <- openTopic_wide[, c("event", "artwork", "card", "trace", openTopic_wide$popup <- NA
openTopic_wide$x.start <- NA
openTopic_wide$x.stop <- NA
openTopic_wide$y.start <- NA
openTopic_wide$y.stop <- NA
openTopic_wide$distance <- NA
openTopic_wide$scale.start <- NA
openTopic_wide$scale.stop <- NA
openTopic_wide$scaleSize <- NA
openTopic_wide$rotation.start <- NA
openTopic_wide$rotation.stop <- NA
openTopic_wide$rotationDegree <- NA
dat_openTopic <- openTopic_wide[, c("event", "artwork", "trace",
"date.start", "date.stop", "date.start", "date.stop",
"time_ms.start", "time_ms.stop", "time_ms.start", "time_ms.stop",
"duration")] "duration", "card", "popup", "x.start",
"y.start", "x.stop", "y.stop",
"distance", "scale.start", "scale.stop",
"scaleSize", "rotation.start",
"rotation.stop", "rotationDegree")]
# TODO: card should have a unique identifier for each artwork # TODO: card should have a unique identifier for each artwork
rm(openTopic_wide, num_start, tmp) rm(openTopic_wide, num_start, tmp)
@ -385,29 +428,47 @@ openPopup_wide$duration <- ifelse(openPopup_wide$duration < 0,
# TODO: How to handle duration < 0 # TODO: How to handle duration < 0
# --> Replace with NA for now... # --> Replace with NA for now...
dat_openPopup <- openPopup_wide[, c("event", "artwork", "popup", "trace", openPopup_wide$card <- NA
openPopup_wide$x.start <- NA
openPopup_wide$x.stop <- NA
openPopup_wide$y.start <- NA
openPopup_wide$y.stop <- NA
openPopup_wide$distance <- NA
openPopup_wide$scale.start <- NA
openPopup_wide$scale.stop <- NA
openPopup_wide$scaleSize <- NA
openPopup_wide$rotation.start <- NA
openPopup_wide$rotation.stop <- NA
openPopup_wide$rotationDegree <- NA
dat_openPopup <- openPopup_wide[, c("event", "artwork", "trace",
"date.start", "date.stop", "date.start", "date.stop",
"time_ms.start", "time_ms.stop", "time_ms.start", "time_ms.stop",
"duration")] "duration", "card", "popup", "x.start",
"y.start", "x.stop", "y.stop",
"distance", "scale.start", "scale.stop",
"scaleSize", "rotation.start",
"rotation.stop", "rotationDegree")]
rm(num_start, openPopup_wide, tmp) rm(num_start, openPopup_wide, tmp)
# Merge all # Merge all
system.time({ # system.time({
dat_all <- merge(dat_trans, dat_flipCard, all = TRUE) # dat_all <- merge(dat_trans, dat_flipCard, all = TRUE)
dat_all <- merge(dat_all, dat_openTopic, all = TRUE) # dat_all <- merge(dat_all, dat_openTopic, all = TRUE)
dat_all <- merge(dat_all, dat_openPopup, all = TRUE) # dat_all <- merge(dat_all, dat_openPopup, all = TRUE)
}) # })
#
# check # # check
nrow(dat_all) == (nrow(dat_trans) + nrow(dat_flipCard) + # nrow(dat_all) == (nrow(dat_trans) + nrow(dat_flipCard) +
nrow(dat_openTopic) + nrow(dat_openPopup)) # nrow(dat_openTopic) + nrow(dat_openPopup))
#
dat_all <- dat_all[order(dat_all$date.start), ] # dat_all <- dat_all[order(dat_all$date.start), ]
rownames(dat_all) <- NULL # rownames(dat_all) <- NULL
#
# TODO: from here on NA... WHY?? # TODO: from here on NA... WHY??
dat_all[19426:19435, ] # dat_all[19426:19435, ]
# TODO: Should card maybe also be filled in for "openPopup"? # TODO: Should card maybe also be filled in for "openPopup"?
@ -423,19 +484,135 @@ dat_all[19426:19435, ]
# TODO: --> same result - but faster. Need it? # TODO: --> same result - but faster. Need it?
# --> Would hate to depend on dplyr... # --> Would hate to depend on dplyr...
#' ## Use `rbind()` instead...
# --> unbeatable in terms of time!
dat_all <- rbind(dat_trans, dat_flipCard, dat_openTopic, dat_openPopup)
# check
nrow(dat_all) == (nrow(dat_trans) + nrow(dat_flipCard) +
nrow(dat_openTopic) + nrow(dat_openPopup))
# remove all events that do not have a `date.start`
dat_all <- dat_all[!is.na(dat_all$date.start), ]
# TODO: Find out how it can be that there is only a `date.stop`
# sort by `start.date`
dat_all <- dat_all[order(dat_all$date.start), ]
rownames(dat_all) <- NULL
ind <- rowSums(is.na(dat_all)) == ncol(dat_all)
any(ind)
dat_all[ind, ]
# --> No rows with only NA, as it should be.
summary(dat_all) # OK, this actually makes a lot of sense :)
#' ## Create case variable
#dat_all$timediff <- as.numeric(dat_all$date.stop - dat_all$date.start)
dat_all$timediff <- as.numeric(diff(c(dat_all$date.start[1], dat_all$date.start)))
hist(dat_all$timediff[dat_all$timediff < 40], breaks = 50)
#' ## Plots # TODO: What is the best choice for the cutoff here? I took 20 secs for now
dat_all$case <- NA
j <- 1
counts <- table(as.Date(dat$date), dat$event) for (i in seq_len(nrow(dat_all))) {
lattice::barchart(counts, auto.key = TRUE) if (dat_all$timediff[i] < 21) {
dat_all$case[i] <- j
} else {
j <- j + 1
dat_all$case[i] <- j
}
}
head(dat_all[, c("event", "artwork", "trace", "date.start", "timediff", "case")], 100)
start_events <- c("Transform start", "Show Info", "ShowPopup", "Artwork/OpenCard") #' ## Add event ID
counts <- table(as.Date(dat$date[dat$event %in% start_events]), dat_all$eventid <- seq_len(nrow(dat_all))
dat$event[dat$event %in% start_events])
lattice::barchart(counts, auto.key = TRUE) dat_all <- dat_all[, c("eventid", "case", "trace", "event", "artwork",
"date.start", "date.stop", "time_ms.start",
"time_ms.stop", "duration", "card", "popup",
"x.start", "y.start", "x.stop", "y.stop",
"distance", "scale.start", "scale.stop",
"scaleSize", "rotation.start", "rotation.stop",
"rotationDegree")]
#' ## Add `trace` numbers for `move` events
# when case and artwork are identical and there is only 1 trace value
# --> assign it to all `move` events for that case and artwork
# when case and artwork are identical and there is more than 1 trace value
# --> assign the `trace` value that was right before this `move` event
# (could, of course, also be after)
cases <- unique(dat_all$case)
aws <- unique(dat_all$artwork)[unique(dat_all$artwork) != "glossar"]
max_trace <- max(dat_all$trace, na.rm = TRUE) + 1
out <- NULL
for (case in cases) {
for (art in aws) {
tmp <- dat_all[dat_all$case == case & dat_all$artwork == art, ]
if (nrow(tmp) != 0) {
if (length(na.omit(unique(tmp$trace))) == 1) {
tmp[tmp$event == "move", "trace"] <- na.omit(unique(tmp$trace))
} else if (length(na.omit(unique(tmp$trace))) > 1) {
for (i in 1:nrow(tmp)) {
if (tmp$event[i] == "move") {
if (i == 1) {
tmp$trace[i] <- na.omit(unique(tmp$trace))[1]
} else {
tmp$trace[i] <- tmp$trace[i - 1]
}
}
}
} else if (all(is.na(tmp$trace))) {
for (i in 1:nrow(tmp)) {
if (tmp$event[i] == "move") {
tmp$trace[i] <- max_trace
}
}
}
max_trace <- max_trace + 1
}
if (nrow(tmp) > 0) {
#print(tmp[, c("case", "event", "trace", "artwork")])
out <- rbind(out, tmp)
}
}
}
# TODO: Get rid of the loops
# --> This takes forever...
#head(out[, c("time_ms.start", "case", "trace", "event", "artwork")], 55)
#head(dat_all[dat_all$artwork %in% "501", c("time_ms.start", "case", "trace", "event", "artwork")], 50)
# identical(dat_all[which(!dat_all$eventid %in% out$eventid), ],
# dat_all[dat_all$artwork == "glossar", ])
# --> TRUE
# put glossar events back in
dat_all <- rbind(out, dat_all[dat_all$artwork == "glossar", ])
dat_all <- dat_all[order(dat_all$date.start), ]
rownames(dat_all) <- NULL
# Make `trace` a consecutive number
dat_all$trace <- as.numeric(as.factor(dat_all$trace))
#' # Export data
write.table(dat_all, "../data/event_logfiles.csv",
sep = ";", quote = FALSE, row.names = FALSE)
# Is `artwork` my case? Or `artwork` per day? Or `artwork` per some other # Is `artwork` my case? Or `artwork` per day? Or `artwork` per some other
@ -451,16 +628,5 @@ lattice::barchart(counts, auto.key = TRUE)
# artwork # artwork
# dat_art <- split(dat, dat$artwork) # dat_art <- split(dat, dat$artwork)
## --> Maybe need it at some point?
#' # Problems
#' * Opening and closing of events cannot be identified unambiguously; it
#' can happen that the wrong tags have been put together (e.g., Transform
#' start and Transform stop); therefore, durations etc. are only heuristic
# TODO: Add a case identifier based on timestamps
# --> needs to be done on "raw data". Is it possible? Something seems
# seriously wrong with `time_ms`
# TODO: Write function for closing events # TODO: Write function for closing events

115
code/03_modeling.R Normal file
View File

@ -0,0 +1,115 @@
#' ---
#' title: "Modelling log files with Process Mining"
#' author: "Nora Wickelmaier"
#' date: "`r Sys.Date()`"
#' output:
#' html_document:
#' toc: true
#' toc_float: true
#' pdf_document:
#' toc: true
#' number_sections: true
#' geometry: margin = 2.5cm
#' ---
# setwd("C:/Users/nwickelmaier/Nextcloud/Documents/MDS/2023ss/60100_master_thesis/code")
#' # Read data
dat <- read.table("../data/event_logfiles.csv", sep = ";", header = TRUE)
dat$date.start <- as.POSIXct(dat$date.start)
dat$date.stop <- as.POSIXct(dat$date.stop)
#' # Creating event logs
library(bupaverse)
names(dat)[6:7] <- c("start", "complete")
table(table(dat$start))
# --> hmm...
summary(aggregate(duration ~ trace, dat, mean))
alog <- activitylog(dat,
case_id = "trace",
activity_id = "event",
#resource_id = "case",
resource_id = "artwork",
timestamps = c("start", "complete"))
# --> have not understood, yet, which ist what...
process_map(alog)
process_map(alog, frequency("relative"))
process_map(alog, frequency("relative_consequent"))
library(processanimateR)
animate_process(to_eventlog(alog))
col_vector <- c("#7FC97F", "#BEAED4", "#FDC086", "#FFFF99", "#386CB0",
"#F0027F", "#BF5B17", "#666666", "#1B9E77", "#D95F02",
"#7570B3", "#E7298A", "#66A61E", "#E6AB02", "#A6761D",
"#666666", "#A6CEE3", "#1F78B4", "#B2DF8A", "#33A02C",
"#FB9A99", "#E31A1C", "#FDBF6F", "#FF7F00", "#CAB2D6",
"#6A3D9A", "#FFFF99", "#B15928", "#FBB4AE", "#B3CDE3",
"#CCEBC5", "#DECBE4", "#FED9A6", "#FFFFCC", "#E5D8BD",
"#FDDAEC", "#F2F2F2", "#B3E2CD", "#FDCDAC", "#CBD5E8",
"#F4CAE4", "#E6F5C9", "#FFF2AE", "#F1E2CC", "#CCCCCC",
"#E41A1C", "#377EB8", "#4DAF4A", "#984EA3", "#FF7F00",
"#FFFF33", "#A65628", "#F781BF", "#999999", "#66C2A5",
"#FC8D62", "#8DA0CB", "#E78AC3", "#A6D854", "#FFD92F",
"#E5C494", "#B3B3B3", "#8DD3C7", "#FFFFB3", "#BEBADA",
"#FB8072", "#80B1D3", "#FDB462", "#B3DE69", "#FCCDE5",
"#D9D9D9")
animate_process(to_eventlog(alog), mode = "relative", jitter = 10, legend = "color",
mapping = token_aes(color = token_scale("artwork",
scale = "ordinal",
range = col_vector)))
elog <- to_eventlog(alog)
animate_process(elog[elog$artwork == "054", ])
animate_process(elog[elog$artwork == "080", ])
animate_process(elog[elog$artwork == "501", ])
process_map(alog[alog$artwork == "054", ])
animate_process(elog[elog$artwork %in% c("080", "054"), ],
mode = "relative", jitter = 10, legend = "color",
mapping = token_aes(color = token_scale("artwork",
scale = "ordinal",
range = c("black", "gray"))))
# --> not sure, yet, how to interpret this...
alog080 <- activitylog(dat[dat$artwork %in% "080", ],
#case_id = "case",
case_id = "trace",
activity_id = "event",
#resource_id = "trace",
resource_id = "case",
timestamps = c("start", "complete"))
process_map(alog080, frequency("relative"))
alog054 <- activitylog(dat[dat$artwork %in% "054", ],
#case_id = "case",
case_id = "trace",
activity_id = "event",
#resource_id = "trace",
resource_id = "case",
timestamps = c("start", "complete"))
process_map(alog054, frequency("relative"))

View File

@ -0,0 +1,125 @@
#' ---
#' title: "Open Questions"
#' author: "Nora Wickelmaier"
#' date: "`r Sys.Date()`"
#' output:
#' html_document:
#' number_sections: true
#' toc: true
#' ---
#+ include = FALSE
# setwd("C:/Users/nwickelmaier/Nextcloud/Documents/MDS/2023ss/60100_master_thesis/code")
dat <- read.table("../data/event_logfiles.csv", sep = ";", header = TRUE)
dat$date.start <- as.POSIXct(dat$date.start)
dat$date.stop <- as.POSIXct(dat$date.stop)
#' This is what the data look like after preprocessing right now
#+ include = FALSE
mat <- as.data.frame(t(sapply(dat, range, na.rm = TRUE)))
names(mat) <- c("min", "max")
mat$min <- round(as.numeric(mat$min), 1)
mat$max <- round(as.numeric(mat$max), 1)
mat$mean <- round(sapply(dat, function(x) mean(x, na.rm = TRUE)), 1)
mat$missings <- sapply(dat, function(x) sum(is.na(x)))
mat <- mat[!(rownames(mat) %in% c("eventid", "case", "trace", "event", "artwork", "card", "popup", "date.start", "date.stop")), ]
#+ echo = FALSE
knitr::kable(mat)
#' This is only the data for 2016! So only about 2 weeks in December.
# Date ranges
range(dat$date.start)
range(dat$date.stop, na.rm = TRUE)
#' # Units of x and y
#' I assume that x and y are pixel $\to$ correct?
#' But they look weird, when plotted. Is it possible that there are
#' outliers? If yes, how? Do we have the true ranges of the display?
par(mfrow = c(1, 2))
plot(y.start ~ x.start, dat)
abline(v = c(0, 3800), h = c(0, 2150), col = "blue", lwd = 2)
plot(y.stop ~ x.stop, dat)
abline(v = c(0, 3800), h = c(0, 2150), col = "blue", lwd = 2)
aggregate(cbind(x.start, x.stop, y.start, y.stop) ~ 1, dat, mean)
#' Looks like the range should be something like $x = [0, 3800]$ and
#' $y = [0, 2150]$. Do we have the starting coordinates for each artwork?
#'
#' # Unit of scale
summary(dat$scaleSize)
#' I thought it would be some kind of scaling factor, but then I would
#' have expected that `scale.start` is always 1 or something.
#'
#' # Unit of rotation
summary(dat$rotationDegree)
#' This looks pretty clear. Should be degree. Anything else to consider
#' here? I am assuming negative means left, but maybe not?
#'
#' # Meaningful unit for "case"
#' I pretty randomly chose `20 sec` based on this plot. I would love a
#' second opinion. `:)`
timediff <- as.numeric(diff(c(dat$date.start[1], dat$date.start)))
hist(timediff[timediff < 40], breaks = 50)
abline(v = 20, col = "red", lwd = 2)
#' This actually works pretty well and lets me assign `trace` values to the
#' moves. But maybe there are other ideas on how to define this?
dat[1:40, c("date.start", "case", "trace", "event", "artwork")]
#' # Problems with `time_ms`
#' What exactly happens, when `time_ms` goes down again? Why does it not go
#' down to 0?
par(mfrow = c(1, 2))
plot(dat$time_ms.start[1:100], type = "b", ylab = "time_ms", xlab = "")
points(dat$time_ms.stop[1:100], type = "b", col = rgb(1, 0, 0, .5))
legend("topleft", c("start", "stop"), lty = 1, col = c("black", "red"))
plot(dat$time_ms.stop[1:100] - dat$time_ms.start[1:100], type = "b",
ylab = "duration", col = rgb(0, 0, 1, .5))
abline(h = 0, lty = 2)
#' For the regular timestamps everything looks fine.
par(mfrow = c(1, 2))
plot(dat$date.stop[1:100], type = "b", ylab = "timestamp", xlab = "",
col = rgb(1, 0, 0, .5))
points(dat$date.start[1:100], type = "b")
legend("topleft", c("start", "stop"), lty = 1, col = c("black", "red"))
plot(dat$date.stop[1:100] - dat$date.start[1:100], type = "b",
ylab = "duration", col = rgb(0, 0, 1, .5))
abline(h = 0, lty = 2)
#+
plot(time_ms.start ~ date.start, dat[1:1000, ], type = "b")
points(time_ms.stop ~ date.stop, dat[1:1000, ], type = "b", col = rgb(1, 0, 0, .3))
#' For `time_ms.stop` this looks even weirder.
#'
#' # After which time interval does the table reset?
#' I cannot see this in the data at all. Or can I? Has this something to do
#' with the weird behavior of `time_ms`?

View File

@ -0,0 +1,111 @@
#' ---
#' title: "Programming input"
#' author: "Nora Wickelmaier"
#' date: "`r Sys.Date()`"
#' output: html_document
#' ---
#+ include = FALSE
# setwd("C:/Users/nwickelmaier/Nextcloud/Documents/MDS/2023ss/60100_master_thesis/code")
#+
dat0 <- read.table("../data/rawdata_logfiles_small.csv", sep = ";", header = TRUE)
dat0$date <- as.POSIXct(dat0$date) # create date object
# Remove irrelevant events
dat <- subset(dat0, !(dat0$event %in% c("Start Application", "Show Application")))
str(dat)
# make data better manageable
tmp <- dat[!dat$event %in% c("Transform start", "Transform stop"), ]
rownames(tmp) <- NULL
#' # Add `trace` variable for closing events
tmp$trace <- NA
last_event <- tmp$event[1]
aws <- unique(tmp$artwork)[unique(tmp$artwork) != "glossar"]
for (art in aws) { # select artwork
for (i in 1:nrow(tmp)) { # go through rows
if (last_event == "Show Info" & tmp$artwork[i] == art) {
tmp$trace[i] <- i
j <- i
} else if (last_event == "Show Front" & tmp$artwork[i] == art) {
tmp$trace[i] <- j
} else if (!(last_event %in% c("Show Info", "Show Front")) &
tmp$artwork[i] == art) {
tmp$trace[i] <- j
}
if (i <= nrow(tmp)) {
last_event <- tmp$event[i + 1]
}
}
}
head(tmp[, c("artwork", "event", "trace")], 50)
#' # Find artwork for glossar entry
glossar_files <- unique(tmp[tmp$artwork == "glossar", "popup"])
# Load lookup table for artworks and glossar files
load("../data/glossar_dict.RData")
lut <- glossar_dict[glossar_dict$glossar_file %in% glossar_files, ]
# Fill in trace variable based on last `Show Info`
for (file in lut$glossar_file) {
artwork_list <- unlist(lut[lut$glossar_file == file, "artwork"])
for (i in seq_len(nrow(tmp))) {
if (tmp$event[i] == "Show Info") {
current_artwork <- tmp[i, "artwork"]
j <- i
k <- i
} else {
current_artwork <- current_artwork
}
if (tmp$event[i] == "Show Front" & tmp$artwork[i] == current_artwork) {
# make sure artwork has not been closed, yet!
k <- i
}
if (tmp$artwork[i] == "glossar" &
(current_artwork %in% artwork_list) &
tmp$popup[i] == file & (j-k == 0)) {
tmp[i, "trace"] <- tmp[j, "trace"]
}
}
}
tmp[tmp$artwork == "glossar", c("artwork", "event", "popup", "trace")]
proportions(table(is.na(tmp$trace[tmp$artwork == "glossar"])))
# --> finds about half of the glossar entries for small data set...
# REMEMBER: It can never be 100% correct, since it is always possible that
# several cards are open and that they link to the same glossar entry
# How many glossar_files are only associated with one artwork?
lut[sapply(lut$artwork, length) == 1, "glossar_file"]
# TODO: Fill in the ones that are associated with one artwork
# --> Can't come up with something -- maybe ask Philipp???
# TODO: How to check if one of the former "Show Infos" is correct
# --> Can't come up with something -- maybe ask Philipp???