It's not really about productivity in the language itself. It's more about the library ecosystem. Numpy and scipy are just so good and nothing really compares.

It's not really about productivity in the language itself. It's more about the library ecosystem. Numpy and scipy are just so good and nothing really compares. I'm not skittish on writing the algorithm I need myself in Haskell. But when I'm prototyping and quickly trying stuff it's just so much faster if I can call a function and be done with it.