Using Tembo's Time Series Stack at Home

The Tembo Time Series stack is now available, but as an author of its underlying extension, I’ve struggled with the same problem I hit any time I contribute to a time-series or database project: I’ve spent more time making and testing this tool than I will spend using it.

Sure, there are public data sets at places like Kaggle, and I like the bike-sharing data set we use in our Time Series guide, but really digging into the data becomes much easier when it’s something of personal interest.

Suddenly it dawned on me: I have a novel IoT use-case of my own… I live in this use-case. My personal hobby of home automation has resulted in a Home Assistant installation with around 350 entities, and I have over a year of statistics!

So now the goal is clear: we’ll be migrating an existing time-series system over to a Tembo Time Series Stack. So readers can follow along, I’m providing a sanitized version of my IoT dataset. Once we have our stack up and running, we’ll cover some interesting time-series queries and finish with some stack benchmarks.

To get started, we’ll need an instance of the Tembo Time Series Stack, so launch one if you haven’t already. We’ll use a GP-4 instance, but the data should fit just fine in the Hobby tier as well.

Leave a Comment