Defining Transformations in Technical Terms

Question and objective

How do technical fields become reliable metrics? A transformation can convert values, standardize fields, or derive a business classification from signals.

Implementation in Bytefabrik

The Script Transformation is available for simple adjustments to the source data. For further processing, use appropriate pipeline elements or custom AI pipelines. The choice of where to implement the processing depends on reusability, complexity, and the required state across multiple events.

Example: A technical status code is assigned to a specific function and supplemented with the system reference. Code 30, for example, stands for "Maintenance." Open full-size image (new tab)

Best Practices

Describe each rule with an input example and the expected output:

Rule	For further clarification, see
Convert milliseconds to seconds	Include the unit in the output field.
Standardize product identifiers	Evaluate leading zeros and case sensitivity from a technical perspective.
Combining Multiple Signals into a Single State	Setting Priority for Conflicting Signals.
Calculating Duration from Start and End	Handling Mapped, Missing End, and Delayed Events.

Check each rule individually before linking multiple steps. For important derivations, keep track of the source events, machine, and time.

What You Should Keep in Mind

A change in data type alone does not convey any business meaning. A numeric product ID remains an identifier and is not a meaningful measurement value.

Unknown states should remain identifiable. Automatically assigning them to the most common state can mask data issues. Additionally, when logic changes, check to determine when old and new results become comparable.

The benefit lies in consistent rules for multiple charts and analyses. Data quality checks ensure that these rules hold true even with real-world inputs.