Went over clustering and helper scripts

This commit is contained in:
Nora Wickelmaier 2024-01-30 11:48:48 +01:00
parent 6ade6444ac
commit 76af291686
5 changed files with 66 additions and 316 deletions

View File

@ -1,9 +1,6 @@
#%reset
import pm4py
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from python_helpers import eval_pm, pn_infos
@ -11,43 +8,21 @@ from python_helpers import eval_pm, pn_infos
dat = pd.read_csv("results/haum/event_logfiles_2024-01-18_09-58-52.csv", sep = ";")
dat = dat[dat["date.start"] < "2020-03-13"]
dat = dat[dat["path"] != 106098] # exclude broken trace
# --> only pre corona (before artworks were updated)
dat = dat[dat["path"] != 106098]
# exclude broken trace
log_path = pm4py.format_dataframe(dat, case_id = "path", activity_key = "event",
timestamp_key = "date.start")
###### Infos for items ######
mdi = pd.DataFrame(columns = ["fitness", "precision", "generalizability",
eval = pd.DataFrame(columns = ["fitness", "precision", "generalizability",
"simplicity", "sound", "narcs", "ntrans",
"nplaces", "nvariants", "mostfreq"])
for item in log_path.item.unique().tolist():
mdi = pd.concat([mdi, pn_infos(log_path, "item", item)])
mdi = mdi.sort_index()
eval = pd.concat([eval, pn_infos(log_path, "item", item)])
eval = eval.sort_index()
# Export
mdi.to_csv("results/haum/pn_infos_items.csv", sep = ";")
# datitem = dat.groupby("item")[["duration", "distance",
# "scaleSize", "rotationDegree"]].mean()
#
# def length_path(data):
# x = data.path
# return len(x.unique())
# def length_case(data):
# x = data.case
# return len(x.unique())
# def length_topic(data):
# x = data.topic.dropna()
# return len(x.unique())
#
# datitem["npaths"] = dat.groupby(["item"]).apply(length_path)
# datitem["ncases"] = dat.groupby(["item"]).apply(length_case)
# datitem["ntopics"] = dat.groupby(["item"]).apply(length_topic)
#
# datitem.index = datitem.index.astype(str).str.rjust(3, "0")
# datitem = datitem.sort_index()
# datitem.index = mdi.index
#
# datitem = pd.concat([mdi, datitem], yaxis = 1)
eval.to_csv("results/haum/pn_infos_items.csv", sep = ";")

View File

@ -1,5 +1,24 @@
# 05_item-clustering.R
#
# content: (1) Read data
# (1.1) Read log event data
# (1.2) Read infos for PM for infos
# (1.3) Extract additional infos for clustering
# (2) Clustering
# (3) Visualization with pictures
#
# input: results/haum/event_logfiles_2024-01-18_09-58-52.csv
# results/haum/pn_infos_items.csv
# output: results/haum/event_logfiles_pre-corona_with-clusters.csv
#
# last mod: 2024-01-30
# setwd("C:/Users/nwickelmaier/Nextcloud/Documents/MDS/2023ss/60100_master_thesis/analysis/code")
library(bupaverse)
library(factoextra)
#--------------- (1) Read data ---------------
#--------------- (1.1) Read log event data ---------------
@ -46,10 +65,8 @@ mat <- dist(df)
hc <- hclust(mat, method = "ward.D2")
library(factoextra)
fviz_dend(hc, cex = 0.5)
datitem$grp <- cutree(hc, k = 6)
grp <- cutree(hc, k = 6)
datitem$grp <- grp
fviz_dend(hc, k = 6,
cex = 0.5,
@ -70,19 +87,25 @@ p
aggregate(cbind(duration, distance, scaleSize , rotationDegree, npaths,
ncases, ntopics) ~ grp, datitem, mean)
datitem$item <- gsub("item_([0-9]{3})", "\\1", row.names(datitem))
datitem$item <- sprintf("%03d",
as.numeric(gsub("item_([0-9]{3})", "\\1", row.names(datitem))))
res <- merge(dat, datitem[, c("item", "grp")], by = "item", all.x = TRUE)
res <- res[order(res$fileId.start, res$date.start, res$timeMs.start), ]
# Look at clusters
vioplot::vioplot(duration ~ grp, res)
vioplot::vioplot(distance ~ grp, res)
vioplot::vioplot(scaleSize ~ grp, res)
vioplot::vioplot(rotationDegree ~ grp, res)
write.table(res,
file = "results/haum/event_logfiles_pre-corona_with-clusters.csv",
sep = ";",
quote = FALSE,
row.names = FALSE)
library(bupaverse)
# DFGs for clusters
res$start <- res$date.start
res$complete <- res$date.stop
@ -95,9 +118,9 @@ for (cluster in sort(unique(res$grp))) {
timestamps = c("start", "complete"))
dfg <- process_map(alog,
type_nodes = frequency("relative"),
type_nodes = frequency("relative", color_scale = "Greys"),
sec_nodes = frequency("absolute"),
type_edges = frequency("relative"),
type_edges = frequency("relative", color_edges = "#FF6900"),
sec_edges = frequency("absolute"),
rankdir = "LR",
render = FALSE)

View File

@ -1,9 +1,5 @@
%reset
import pm4py
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from python_helpers import eval_pm, pn_infos
@ -17,15 +13,14 @@ log_path = pm4py.format_dataframe(dat, case_id = "path", activity_key = "event",
###### Infos for clusters ######
# Merge clusters into data frame
mdc = pd.DataFrame(columns = ["fitness", "precision", "generalizability",
eval = pd.DataFrame(columns = ["fitness", "precision", "generalizability",
"simplicity", "sound", "narcs", "ntrans",
"nplaces", "nvariants", "mostfreq"])
for cluster in log_path.grp.unique().tolist():
mdc = pd.concat([mdc, pn_infos(log_path, "grp", cluster)])
mdc = mdc.sort_index()
eval = pd.concat([eval, pn_infos(log_path, "grp", cluster)])
eval = eval.sort_index()
# Export
mdc.to_csv("results/haum/pn_infos_clusters.csv", sep = ";")
eval.to_csv("results/haum/pn_infos_clusters.csv", sep = ";")
###### Process maps for clusters ######

View File

@ -2,11 +2,13 @@
# setwd("C:/Users/nwickelmaier/Nextcloud/Documents/MDS/2023ss/60100_master_thesis/analysis/code")
library(bupaverse)
#--------------- (1) Look at broken trace ---------------
# Read data
datraw <- read.table("results/haum/raw_logfiles_2024-01-18_09-58-52.csv",
header = TRUE, sep = ";")
dat0 <- read.table("results/haum/event_logfiles_2024-01-18_09-58-52.csv",
datlogs <- read.table("results/haum/event_logfiles_2024-01-18_09-58-52.csv",
colClasses = c("character", "character", "POSIXct",
"POSIXct", "character", "integer",
"numeric", "character", "character",
@ -15,140 +17,17 @@ dat0 <- read.table("results/haum/event_logfiles_2024-01-18_09-58-52.csv",
"character", "character"),
sep = ";", header = TRUE)
dat0$event <- factor(dat0$event, levels = c("move", "flipCard", "openTopic",
"openPopup"))
artwork <- "176"
fileId <- c('2017_06_16-13_49_00.log', '2017_06_16-13_59_00.log')
path <- 106098
dat0$weekdays <- factor(weekdays(dat0$date.start),
levels = c("Montag", "Dienstag", "Mittwoch",
"Donnerstag", "Freitag", "Samstag",
"Sonntag"),
labels = c("Monday", "Tuesday", "Wednesday",
"Thursday", "Friday", "Saturday",
"Sunday"))
datraw[datraw$item == artwork & datraw$fileId %in% fileId, ]
# Select data pre Corona
dat <- dat0[as.Date(dat0$date.start) < "2020-03-13", ]
dat <- dat[dat$path != 106098, ]
datlogs[datlogs$path == path, ]
#--------------- (2) Function to find broken traces ---------------
table(table(dat$start))
table(dat$event)
proportions(table(dat$event))
dat_dur <- aggregate(duration ~ item, dat, mean)
barplot(duration - mean(dat_dur$duration) ~ item, dat_dur, col = "#434F4F",
las = 3)
# Investigate paths (will separate items and give clusters of artworks!)
length(unique(dat$path))
# DFGs per Cluster
dat$start <- dat$date.start
dat$complete <- dat$date.stop
summary(aggregate(duration ~ path, dat, mean))
alog <- activitylog(dat,
case_id = "path",
activity_id = "event",
resource_id = "item",
timestamps = c("start", "complete"))
process_map(alog,
type_nodes = frequency("absolute"),
sec_nodes = frequency("relative"),
type_edges = frequency("absolute"),
sec_edges = frequency("relative"),
rankdir = "LR")
### Separate for items
datitem <- aggregate(cbind(duration, distance, scaleSize, rotationDegree) ~
item, dat, function(x) mean(x, na.rm = TRUE), na.action = NULL)
datitem$npaths <- aggregate(path ~ item, dat,
function(x) length(unique(x)),
na.action = NULL)$path
datitem$ncases <- aggregate(case ~ item, dat,
function(x) length(unique(x)),
na.action = NULL)$case
datitem$ntopics <- aggregate(topic ~ item, dat,
function(x) ifelse(all(is.na(x)), NA, length(unique(na.omit(x)))),
na.action = NULL)$topic
set.seed(1211)
nclusters <- 6
k1 <- kmeans(datitem[, -1], nclusters)
#colors <- c("#3CB4DC", "#78004B", "#91C86E", "#FF6900")
colors <- palette.colors(palette = "Okabe-Ito")
xy <- cmdscale(dist(datitem[, -1]))
plot(xy, type = "n")
text(xy[,1], xy[,2], datitem$item, col = colors[k1$cluster])
legend("topright", paste("Cluster", 1:nclusters), col = colors, lty = 1)
## Scree plot
ks <- 1:10
sse <- NULL
for (k in ks) sse <- c(sse, kmeans(datitem[, -1], k)$tot.withinss)
plot(sse ~ ks, type = "l")
datitem$cluster <- k1$cluster
datitem_agg <- aggregate(. ~ cluster, datitem[, -1], mean)
dat_cl <- merge(dat, datitem[, c("item", "cluster")], by = "item", all.x = TRUE)
dat_cl <- dat_cl[order(dat_cl$fileId.start, dat_cl$date.start, dat_cl$timeMs.start), ]
write.table(dat_cl, "results/haum/event_logfiles_with-clusters_kmeans.csv",
sep = ";", row.names = FALSE)
vioplot::vioplot(datitem$duration)
vioplot::vioplot(duration ~ item, dat, las = 3)
vioplot::vioplot(duration ~ cluster, dat_cl)
vioplot::vioplot(distance ~ cluster, dat_cl)
vioplot::vioplot(scaleSize ~ cluster, dat_cl)
vioplot::vioplot(rotationDegree ~ cluster, dat_cl)
for (cluster in sort(unique(dat_cl$cluster))) {
alog <- activitylog(dat_cl[dat_cl$cluster == cluster, ],
case_id = "path",
activity_id = "event",
resource_id = "item",
timestamps = c("start", "complete"))
dfg <- process_map(alog,
type_nodes = frequency("relative"),
sec_nodes = frequency("absolute"),
type_edges = frequency("relative"),
sec_edges = frequency("absolute"),
rankdir = "LR",
render = FALSE)
export_map(dfg,
file_name = paste0("results/processmaps/dfg_cluster", cluster, "_R.pdf"),
file_type = "pdf",
title = paste("DFG Cluster", cluster))
}
tmp <- dat[dat$event != "move", ]
tmp <- datlogs[datlogs$event != "move", ]
check_traces <- function(data) {
@ -170,127 +49,5 @@ check_traces <- function(data) {
check <- check_traces(tmp)
sum(check$check)
alog <- activitylog(dat,
case_id = "case",
activity_id = "item",
resource_id = "path",
timestamps = c("start", "complete"))
process_map(alog,
type_nodes = frequency("absolute"),
sec_nodes = frequency("relative"),
type_edges = frequency("absolute"),
sec_edges = frequency("relative"),
rankdir = "LR")
datcase <- dat[!duplicated(dat[, c("case", "path", "item")]),
c("case", "path", "event", "item")]
datcase$duration <- aggregate(duration ~ path, dat,
function(x) mean(x, na.rm = TRUE), na.action = NULL)$duration
datcase$distance <- aggregate(distance ~ path, dat,
function(x) mean(x, na.rm = TRUE), na.action = NULL)$distance
datcase$scaleSize <- aggregate(scaleSize ~ path, dat,
function(x) mean(x, na.rm = TRUE), na.action = NULL)$scaleSize
datcase$rotationDegree <- aggregate(rotationDegree ~ path, dat,
function(x) mean(x, na.rm = TRUE), na.action = NULL)$rotationDegree
# datcase$ntopics <- aggregate(topic ~ path, dat,
# function(x) ifelse(all(is.na(x)), NA, length(unique(na.omit(x)))),
# na.action = NULL)$topic
datcase$move <- ifelse(datcase$event == "move", 1, 0)
# paths that start with move
for (item in sort(unique(datcase$item))) {
datcase[paste0("item_", item)] <- ifelse(datcase$item == item, 1, 0)
}
mat <- na.omit(datcase[, -c(1:4)])
set.seed(1610)
nclusters <- 6
k1 <- kmeans(mat, nclusters)
#colors <- c("#3CB4DC", "#78004B", "#91C86E", "#FF6900")
colors <- palette.colors(palette = "Okabe-Ito")[1:nclusters]
library(distances)
mat_dist <- distances(mat)
xy <- cmdscale(mat_dist)
plot(xy, type = "n")
text(xy[,1], xy[,2], datcase$path, col = colors[k1$cluster])
legend("topright", paste("Cluster", 1:nclusters), col = colors, lty = 1)
## Scree plot
ks <- 1:10
sse <- NULL
for (k in ks) sse <- c(sse, kmeans(datitem[, -1], k)$tot.withinss)
plot(sse ~ ks, type = "l")
alog <- activitylog(datcase,
case_id = "case",
activity_id = "item",
resource_id = "path",
timestamps = c("start", "complete"))
process_map(alog,
type_nodes = frequency("relative"),
sec_nodes = frequency("absolute"),
type_edges = frequency("relative"),
sec_edges = frequency("absolute"),
rankdir = "LR")
# setwd("C:/Users/nwickelmaier/Nextcloud/Documents/MDS/2023ss/60100_master_thesis/analysis/code")
datraw <- read.table("results/haum/raw_logfiles_2024-01-18_09-58-52.csv",
header = TRUE, sep = ";")
# Read data
datlogs <- read.table("results/haum/event_logfiles_2024-01-18_09-58-52.csv",
colClasses = c("character", "character", "POSIXct",
"POSIXct", "character", "integer",
"numeric", "character", "character",
rep("numeric", 3), "character",
"character", rep("numeric", 11),
"character", "character"),
sep = ";", header = TRUE)
datlogs <- datlogs[order(datlogs$fileId.start, datlogs$date.start, datlogs$timeMs.start), ]
artwork <- "176"
fileId <- c('2017_06_16-13_49_00.log', '2017_06_16-13_59_00.log')
path <- 106098
datraw[datraw$item == artwork & datraw$fileId %in% fileId, ]
datlogs[datlogs$path == path, ]
check[check$check, ]

View File

@ -19,7 +19,7 @@ def pn_infos(log, colname, filter):
net, im, fm = pm4py.discover_petri_net_inductive(filtered_log)
eval = eval_append(log, net, im, fm)
eval = eval_append(filtered_log, net, im, fm)
eval.index = [str(filter).zfill(3)]
return eval