The Kronts Lab

Practical guides for engineers and data scientists working with industrial time series data

March 3, 2026

Simulating Real-World Shocks: How to Generate Transient Events in Synthetic Time Series

Industrial sensors record critical moments, earthquakes, impacts, valve closures, where sudden transients reveal hidden physics. Yet traditional synthetic data generation treats these as afterthoughts rather than core features, leaving engineers with

Multiple correlated time series image with genAI

Feb. 2, 2026

AI-Assisted Synthetic Time Series: Why Parametrized, Not Code Generation

PHOENIX follows a parametrized generation architecture. Rather than having the LLM generate Python code that gets executed—a tempting approach at first glance—PHOENIX interprets natural language descriptions and translates them into a structured JSON

Feb. 25, 2026

Multichannel Correlated Time Series

When building sensor data pipelines, real data feels like the obvious choice. In practice, it’s harder than it looks, physically constrained synthetic data solves that gap.

Feb. 19, 2026

PHOENIX: Simple and Efficient Synthetic Time Series Generator App

Create and save synthetic time series (time stamped) for testing your data science workflows.