Your question is Pandas and NumPy Performance. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
Explain why NumPy can be much faster than pure Python for numerical computation, and describe how you would handle a Pandas DataFrame that has duplicate column names.