Ask HN: Rust Viable for Data Analytics?
6 comments
Clickable link: https://areweguiyet.com/
Rust still has some key pieces missing, but looks promising, see: https://github.com/wiseaidev/rust-data-analysis
F# has a very decent data community: https://fsharp.org/guides/data-science and https://fslab.org
And obviously Julia is also something to consider. It depends on the compromises you want to make and what you define as data analytics.
F# has a very decent data community: https://fsharp.org/guides/data-science and https://fslab.org
And obviously Julia is also something to consider. It depends on the compromises you want to make and what you define as data analytics.
Thank you! Although I'm not a huge fan of REPL and notebooks, the rust-data-analysis repo might give me at least a starting point.
Currently, I'm not really keen to learn another new language, as I just got started with rust, and want to start using it a bit more.
Currently, I'm not really keen to learn another new language, as I just got started with rust, and want to start using it a bit more.
Rust is great for well defined tasks that are not too exploratory and/or will be ran many times, so performance matters. There are quite a few interesting statistical inference tools being built on Rust right now. Julia is fast, but it still has some memory usage problems in my experience.
One area I am interested in where Rust could make a lot of sense is edge computing & ML. Imagine a model such as a Kalman filter in a little robot. Here, Rust's low footprint and latency would be advantageous.
One area I am interested in where Rust could make a lot of sense is edge computing & ML. Imagine a model such as a Kalman filter in a little robot. Here, Rust's low footprint and latency would be advantageous.
No, it's way too rigid to iterate quickly and work interactively. And it's not a "is the language there yet" - it's not supposed to be there at all, it targets completely different use cases.
Not a whole answer, but maybe worth mentioning that Polars is implemented in rust and you can directly use it via it's Rust crate. https://docs.rs/polars/latest/polars/
But I've started to use rust more and more in the last few weeks and really start to like it.
Does anyone have experience doing data analysis with rust, and would you recommend it over python?
And are there any resources like https://areweguiyet.com/ but for data analysis?