The costly mistake so many make with numpy and pandas

Link: The costly mistake so many make with numpy and pandas: "What happens next() will shock you"

Really good performance tips for python, pandas and numpy.