231 points by quinn 1 year ago flag hide 8 comments
newbie_data_scientist 1 year ago next
Great post! I've been using Python for data science work, but I'm intrigued by Rust's potential. Mind sharing what specific data science tasks Rust excels at?
shifted_to_rust 1 year ago next
Definitely! Rust is great for data processing tasks that require heavy numeric computations. Its performance is similar to C++, but with a safer and more ergonomic syntax. Rust's memory safety and zero-cost abstractions are plus points for large-scale data processing.
precise_engineer 1 year ago prev next
What about Python libraries for data analysis and visualization? Do you use Rust libraries or interface with Python ones?
shifted_to_rust 1 year ago next
Currently, Rust doesn't have a mature ecosystem compared to Python. However, there are some promising projects such as 'ndarray' for numeric arrays, 'statrs' for statistical functions, and 'gnuplot' for plotting. Rust can also use Python libraries via the 'calloop-rg' or 'pyo3' crates.
language_enthusiast 1 year ago prev next
Does the learning curve for Rust impact your productivity?
shifted_to_rust 1 year ago next
Yes, Rust's learning curve is steep, and there is a productivity cost involved. However, investing time in learning Rust has long-term benefits in terms of performance and safety. It took me about a month to become proficient with Rust for data science tasks.
master_of_none 1 year ago prev next
How does Rust compare to Julia for data science?
shifted_to_rust 1 year ago next
Julia has a more mature ecosystem for data science, with a focus on statistical modeling and machine learning. Rust, on the other hand, excels in performance and memory safety, making it suitable when speed and reliability are crucial.