For someone new to time series analysis, how did you choose these particular algorithms? Are they standard in the field, or more of a personal selection?
Most of the algorithms in augurs were chosen to solve problems we've had at Grafana, which tend to require a solution that doesn't require tweaking too many parameters and deals with higher frequency series than many other time series algorithms are designed to deal with. For example, the DBSCAN clustering algorithm works without having to choose the number of clusters, and MSTL/Prophet work with multiple seasonalities and sub-daily data.
The other criteria is that they needed to be fast and cheap, which ruled out many of the deep learning/neural net based models, although I'd still like to try some foundation models using Burn or some other Rust deep learning framework!
Yes, in fact this is exactly what it was written for! If you're writing a Grafana app plugin (using Grafana Scenes, a library to help write dashboard-like experiences in Grafana apps) you can use Scenes ML to add outlier detection, forecasting or changepoint detection pretty easily: https://grafana.com/developers/scenes/scenes-ml/outlier-dete....
Readit News
Show HN: Augurs, a time series toolkit for Rust - https://news.ycombinator.com/item?id=42184386 - Nov 2024 (1 comment)
- Repo: https://github.com/grafana/augurs
- Docs: https://docs.augu.rs/
- Python library: https://pypi.org/project/augurs/
- npm library: https://www.npmjs.com/package/@bsull/augurs
The other criteria is that they needed to be fast and cheap, which ruled out many of the deep learning/neural net based models, although I'd still like to try some foundation models using Burn or some other Rust deep learning framework!