In NiklasCarlos1994/batchanalytics: Functions That Represent Metrics For Analysing Batching Behavior

library(batchanalytics)
library(bupaR)
library(bamalog)
library(tidyr) # warum brauche ich das noch für gather aber die anderen package(bama) nicht mehr da imported mit package batchAanalytics---

Load sample eventlog

csv = system.file("exdata", "sample_data_1.csv", package = "batchanalytics")

csv_log = read.csv(csv)

task_log <- csv_log

detect batching behaviour and create a task log

add documentation why seq_list

result_log <- my_detect_batching(task_log)

#code refacor df_list
#1 ersetzen

#df_lists <- get_batching_df_logs(result_log)

get_batching_df_logs(result_log)

batch processing metrics

# transformation to event_logs
#2 unterer code ersetzten

transform_df_to_event_log()

show process model with batching behaviour

# create this as new method -> aber auch daran denken das "no batching" eine option ist 


show_batching_in_process_map()



#identfy which activites are performed with batching behaviour and if possible highlight in output graph or give an additional text output

Cycle Time

The throughput time is the time form the very first event to the last event of a case. The levels at which it can be computed are log, trace, or case.

(overall cycle time, see pic bupR waiting time included)

#evtl nicht nögig


#cycle time function without bupaR


#view github implementation of bupaR 
# compare to niels function implementation

#result_log %>% group_by("case_id") %>% 

# create this as new method -> aber auch daran denken das "no batching" eine option ist 




#when writing method consider that not always all batch types are included and also no batching possible -> when plotting and creating boxplot create a generic approach#


# für modularität keep event log cration seperate from other functions 

#evtl cycle time selber funktion ohne event log schreiben -> von jeden case start - endzeit (groubby case , start - end , -> mean() an niels orientieren) ähnlich für processing time -< vorher #ABER data spliten in "nobatching " vs andere batch types


compare_throughput_time()

processing time

processing time: the sum of the duration of all activity instances[bupaR]

# create this as new method -> aber auch daran denken das "no batching" eine option ist 
# modularer aufbau der funktionen


compare_processing_time() # warum global env und nicht function

idle time

idle time: the time when no activity instance is active

compare_idle_time()

Metric cycle time efficiency

Ratio of overall processing time relative to the overall cycle time

#cycle time efficiency


# create this as new method -> aber auch daran denken das "no batching" eine option ist 
# modularer aufbau der funktionen



# processing time / throughput time


cycle_time_efficiency()

metric batch frequency

This metrics shows .....

#test batch frequency



metric_frequency(result_log, "B", "WorkerB_#0", "concurrent", FALSE )

#test batch size
#metric_frequency(result_log, "Registration", "r1", "concurrent", FALSE )

metric_batch_size(result_log, "B", "WorkerB_#0", "concurrent", exclude_singletons = TRUE)

# Metric - Activity duration
metric_activity_duration(result_log, "B", "WorkerB_#0", "concurrent")

#create event log for bupaR

  #result_log %>% eventlog( case_id = "case_id", activity_id = "activity")

NiklasCarlos1994/batchanalytics documentation built on Dec. 17, 2021, 5:25 a.m.