dljvandenberg / processminer

Process Mining Exploration Tool

ProcessMiner

Description

ProcessMiner is a simple web-based process mining tool for exploration (and potentially prediction).

It was created by Dennis van den Berg and uses the bupaR process mining library in R, with a UI built in R Shiny.

Its current status is: experimental

Features

The web interface allows the user to:

• Upload an eventlog and mark relevant fields (case_id, timestamp, activity)
• Alternatively, select example eventlogs
• View the raw data
• View summary statistics (such as number of cases, events, activities, distribution of throughput times)
• Interactively generate animated process flows
• Generate timeline views of events

We will consider adding prediction models for throughput times, success/failure per case, etc in future releases.

Maintainer

This git repository is maintained by Dennis van den Berg (https://github.com/dljvandenberg).

processminer's Issues

Refactoring: convert code into modules

• Summary statistics
• Process flow
• Timeline view
• Table view
• Data load

Dynamic size for timeline view box

Scale size of timeline_box and plotlydottedchart in javascript part

Filter on min nr of traces in flow diagram

filter_trace_frequency(interval = c(input$min_trace_frequency, NA))

Update and test Dockerfile

Troubleshoot color by time error in process flow after deployment on shinyapps.io

An error has occurred. Check your logs or contact the app author for clarification.

Deploy app via shinyapps.io

First version deployed and accessible at: https://dljvandenberg.shinyapps.io/processminer/

First version of RShiny app

Add license file

MIT license

Process flow: show case details for selected case

Show table with case details when case is clicked in diagram

Remove orientation and duration options from process flow section

Fix token size by attribute in process flow

Check for correct data type of attribute

Fix process flow error

"Error in if: missing value where TRUE/FALSE needed"

Make menuItems conditional on data load

When eventlog has not been loaded yet, don't show menu tabs that depend on eventlog.

Remove data-specific configuration during data load

Fix: Error: no applicable method for 'dotted_chart' applied to an object of class "NULL"

On Timeline View tab, when no dataset has been loaded.

Update README

Use shinydashboard layout

Menu items:

• Data selection
• Raw data
• Summary statistics
• Process flow
• Timeline view
• About ProcessMiner

Add time filter option to flow diagram section

edeaR::filter_time_period(input$time_period)

Load example eventlog data

Load example eventlog from eventdataR package.

Create process view

Process view using processanimatR

Add buttons with instruction text

• Example datasets
• Data upload
• Summary statistics
• Process flow - settings
• Process flow - diagram
• Events timeline

Improve color scale in process flow

• Check data type
• Choose better color scale

Link/embed ProcessMiner on personal web page

Troubleshoot ssl error during deployment to shinyapps.io

Tweak process map layout

• Use red selectors (that are better visible)
• Move/remove less important dropdown menus
• Add option for default vs performance view (including throughputs)

Fix: Error: no applicable method for 'timestamp' applied to an object of class "NULL"

Visible in settings box on process flow tab, when no dataset has been loaded.

Add subMenuItem to load example dataset(s)

Remove DEBUG messages

Add tab with summary statistics

• Number of cases
• Number of events
• Number of cases per activity
• Number of events per activity
• Throughput times per case

Bigger process flow diagram

Use color bins for throughput times in process map

See https://bupaverse.github.io/processanimateR/articles/use-timestamp-change-token-aes.html

Build Docker image

Create Dockerfile of Rshiny app and build docker image

Fix: Warning: Error in if: argument is of length zero

Stack trace

Warning: Error in if: argument is of length zero
99: func [R/functions.R#311]
86: origRenderFunc
85: output$process
5: runApp
3: print.shiny.appobj
1: source

Preprocess new eventlog data

Layout: add colored titles to boxes

Add upload button for new eventlog data

Fix missing spinner

Remove big example datasets that cause plotting error in process flow

For example: hospital dataset

Show patients dataset as default

Handle color scale for datetime variables

Support for csv file upload

At the moment only Excel files supported.

Issue closed because reading csv's is prone to errors when reading date/datetime format

Add timeline view plot

Using processmapR::plotly_dotted_chart

Refactoring: split into ui.R and server.R

Redeploy app

Add tabsetpanel with data tab

Write down lessons learned so far

• Process mining skills: bupaR, edeaR, processmapR, processanimateR, ..
• R Shiny skills: file upload, shiny dashboard, reactive values, button clicks, modules
• RColorBrewer
• Docker: image building