November 25, 2025
Popcorn ready: language wars ignite
Python is not a great language for data science. Part 1: The experience
R fans cheer, Python loyalists eye-roll, Clojure crashes the party
TLDR: A lab lead says Python is fine but not great for everyday data tasks, praising R’s smoother experience. Comments erupt: pro‑R design love, pushback that the article withholds proof until Part 2, and a Clojure cameo — all debating whether Python’s crown is hype or history.
The author just touched the hot stove: Python is “good” but not “great” for everyday data work like cleaning, charting, and stats — deep learning excluded. In the comments, the fireworks start. “Python’s dominance is a historical accident,” cheers lenerdenator, while R devotees credit community heroes (Di Cook, Hadley Wickham, Yihui Xie) for crafting an ecosystem that feels effortless. paulfharrison adds that R’s design lets packages like the tidyverse feel almost built into the language.
Not everyone’s impressed. forgotpwd16 snarks that Part 1 “dodges its own thesis,” since the real takedown is postponed to a sequel. yeahwhatever10 rolls in with “A little late for this,” suggesting the language war meme has been rerun a thousand times. Then a curveball: iLemming pitches Clojure as a data darling — with a wink that you can still use Python libraries from it.
The meta-debate? If Python needs special add-ons, so does R — so maybe the fight is more about vibe than capability. Commenters riff on the author’s “use the hammer you’ve got” analogy, turning it into jokes about opening beers with code. Want the receipts? The author promises more in Part 2. Popcorn, meet pipelines.
Key Points
- •The author argues Python is overused as the default language for data science and is not great for many non–deep learning tasks.
- •Python is acknowledged as strong for deep learning, with PyTorch cited as an industry standard; deep learning is excluded from the critique.
- •The author often prefers R over Python for various data science tasks such as wrangling, EDA, visualization, and statistical modeling.
- •The perspective is informed by over two decades running a computational biology lab with about thirty trainees and a permissive tool-choice policy.
- •The author reports practical frustration with the effort required for routine, iterative data analysis workflows in Python.