Hi @ThiamLee , How are you doing today?

As per my experience, I’d suggest giving Genie Code a try, especially if you’re already working in Databricks.

I’ve been using Genie Code while building a few of my own Data + AI apps and POCs, and what I like most is that it feels less like a code generator and more like a development partner. I can give it the dataset and explain what I’m trying to achieve, and it helps me explore the data, write the Python or PySpark code, test different approaches, debug issues and keep improving the solution.

For your time-series use case, I’d start by asking Genie Code to analyze the existing dataset first: frequency, seasonality, trends, correlations, distributions and anomalies. Then ask it to build a synthetic data generation approach that preserves those patterns. You can even iterate with simple instructions like “generate 10x more records,” “preserve weekly seasonality,” “introduce realistic anomalies,” or “compare the synthetic data against the original.”

I’ve found this kind of iterative workflow very useful while building my apps because you can go from an idea to an actual working experiment much faster without having to write everything from scratch.

Worth trying before adding another standalone tool to the stack. Hope this helps!