I can't seem to find a library for computing canonical correlation in Spark. Canonical correlation analysis (CCA) is common in other data science-focused systems.
This exists in R's base and in Python in Scikit. Julia's Multivariate.jl has it but is quite broken.
