Perceptif Process Mining Framework
This article is an introduction to Perceptif Analysis Methodology to help you gain insights from your data.
📘 Methodology
Perceptif Framework for Process Mining Analysis
The framework above can be iterated as much as needed to get the full insights from the data. There are three main steps in Perceptif Framework for process mining analysis, as follows.
Planning
The project's expectations and goals are intended to be established during the planning phase. Choosing which process needs to be examined depends on the quality of the data and how adaptable the process is. Processes without high-quality data will not produce excellent analytical findings since process mining depends on solid event logs. To determine the expected results, project goals, and main questions must then be established. The following are the details for each planning step.
Formulate Questions
Arrange questions that are keen to be answered. Questions can be related to actual process conditions, process performance, as well as possible bottlenecks, compliance checks, and others
Get the context and knowledge about the data and process from process owners
Each repetition of the analysis can yield more precise questions. More specific questions frequently surface as the analysis progresses. Avoid getting bogged down in minute details and always consider the insights as a part of the overall process.
Prepare and Import Data
Make sure the data is suitable for analysis (csv, xlsx, xes)
Make sure the connection to third party stream works
Make sure the data has all the required attributes and are assigned correctly
Analysis
In order to respond to the main questions posed in the planning phase, process mining techniques are applied to the event log during the analysis step. Process discovery, compliance checking, and performance analytics are the four methods/steps covered by the analysis. Each stage in this phase can be done many times and in parallel. The following are the details for each analysis step.
Discover Process Maps
Flow Analysis
Check how the actual process plays out
Check which activities start the process and which end the process
Check all possible arcs (connection) and nodes (activities)
Variant Analysis
Check the top 10 most common variants
Use a variant analyzer to get a helicopter view of how the process played out
Filter analysis
Use Attribute selection filters to retain/remove cases based on selected attributes in the data
Use activity selection filters to retain/remove cases that are flow-through/do not flow through/start with/end with certain activities
Use bottleneck selection filters to retain/remove cases that are followed anytime by/directly followed by/not directly followed by/never followed by certain activities
Use rework filters to retain cases which rework process
Check Compliance
Throughput time analysis
Use throughput time filters to retain cases that are active in a specific duration
Check the median/average duration of each process and the waiting time between each process
Compare the process maps with the ideal process model
Import ideal process model (BPMN) to the whiteboard
Compare the actual process maps to the process model
Compare the throughput time with the SLA time
Performance Analysis
Bottleneck analysis
Check the throughput time between activities and identify which activities took the longest
Check the activities sequence that always exists in variants
Dashboard and visualization
Visualize the data and case attributes with widgets
Build dashboards to monitor the process and gain insights
Connect to third-party BI tools for analysis with Clickhouse Connection
Improvement
The analysis's findings are then used to provide improvement suggestions that are in keeping with the objectives. Possible process improvements including automation can be implemented with Perceptif.
Automate for improvement
Automate processes to improve efficiency based on recommendation
Monitoring and evaluation by setting up alarms for inefficiency, etc