Update: Automating Time Series Exploration with tseda 📈

A while back, I shared tseda, a tool designed to help you make sense of high-frequency business metrics (like hourly conversion rates or service windows).
Since then, I’ve been working on making the transition from "collecting data" to "understanding data" even faster.

What’s New in tseda?

• Automatic Window Management: You no longer have to guess your window sizes. The tool now handles automatic window size assignment and refinement, finding the "signal" in your data without the trial and error.
• Notebook Parity: You can now move seamlessly between the tool and Jupyter notebooks. Keep your flow state intact while switching from visual exploration to deep-dive coding.

Why use it?

If you have data at an hourly or greater cadence, you’re likely looking for two things: Forecasting and Anomaly Detection. tseda is built to help you build better apps by actually understanding the underlying patterns of those metrics.

Get Started (or Catch Up):

• New README & Docs: github.com/rajivsam/tseda
• User Guide: Step-by-step instructions
• Video Overview: AI-generated summary

I’m looking for feedback from anyone monitoring metrics at a high cadence. How are you currently handling window refinements? Let’s discuss in the comments!

Would you like me to tailor the technical highlights to focus more on the Markov analysis or the specific Python libraries you used?