• Project Introduction: Announced augurs, a Rust time series toolkit with functionality like forecasting (ETS, MSTL, Prophet), outlier detection (DBSCAN, MAD), dynamic time warping (DTW), time series clustering (DBSCAN), and seasonality detection (periodograms). It's a work in progress and can be demoed at demo.augu.rs.
• What is a Time Series: A sequence of data points indexed by time, like temperature readings taken hourly for a year.
• Why Use This: Time series data is common, and while there are many tools in Python or R, there are few in Rust. It fills this gap and provides bindings for Javascript and Python. Real-world use cases include projecting data for planning, detecting anomalies and outliers, and finding similar series groups. It also feels faster than other implementations based on anecdotal evidence and some benchmarks.
• Name Explanation: "Augur" means "to predict", which was the initial intention of the project. The augu.rs domain was available.
• How to Use: Best to look at the documentation. The Rust API docs are thorough, but the Javascript and Python bindings need more work. Here's a quick example of using augurs for in-sample forecasting with the Prophet model in Rust code.
• Contribution: The GitHub repository is the best place to start for contributions. Feature requests and bug reports are welcome.