Optimizing Holding Data with Rolling Means: A Comparison of Two Methods in Python