PROCESS MINING
In the 420+ approach, Process Mining compares suitable recorded batch histories. Recurring differences in quality, results, or resource consumption provide leads for targeted investigation. Which conclusions are possible depends on the data's scope, preparation, and the methods used.
Process histories become comparable.
DATA CAPTURE CLOSE TO THE PROCESS
A comparison starts with a specific question: Which differences between the histories could be relevant to the result? This requires, for example, material, SOP version, measurements, duration, and result. Missing or ambiguous assignments limit the analysis.
Material, source batch, tool, equipment, room, additions, measurements, SOP version, reviews, duration, and result remain connected to each history.
Process Events record individual actions and observations with their respective times and process context.
Object-Centric Process Mining examines interacting object histories, such as those of a batch, sample, and test order. Choosing this perspective changes which relationships and waiting times become visible.
The structured foundation is created during work. Additional data sources, assignments, and quality checks depend on the specific analysis objective. The articles below explore exchange formats, model discovery, and conformance checking.
Capture close to the process creates the starting point for analysis.
STANDARDIZED IMAGE CAPTURE
In the platform concept, images can be connected to product, batch, and time at defined work steps. Comparable capture additionally requires suitable conditions, such as perspective, lighting, and a defined subject. Process context alone does not standardize the image.
The image is therefore not isolated in a folder but sits at the exact point in the history where it was created—together with material, conditions, measurements, and result.
This connection makes images analyzable alongside the other process data. Changes can be compared over time and included in data and image analyses. On the same foundation, images could later also be captured directly during work—for example, through smart glasses.
The value lies not in the image alone, but in its unambiguous place in the history.
COMPARISON AND DISTILLATION
Suitable histories are selected for comparison using defined criteria. Differences can justify an investigation question but do not yet explain a cause. In the described 420+ workflow, the selection is passed to Process Intelligence for further substantive assessment.
The filtered selection is passed to Process Intelligence.
repeatedly occurred together with a higher result in comparable histories.
Many histories reveal a few differences that deserve the next closer look.
EXPLORE THE KNOWLEDGE BASE
Further articles explain the data models and analysis methods behind actual process histories.
NEXT STEP
The selected relationships form the foundation for the next system step.
