How do you decide whether to use Scala or Python (PySpark) for a large-scale data processing pipeline?
