Hands-On Data Analysis with Pandas

Z-score

Remember, from Chapter 1, Introduction to Data Analysis, that we can also calculate Z-scores and flag IP addresses a given number of standard deviations from the mean. The pct_change_threshold() function we wrote earlier won't help us as is, since we aren't just comparing to the baseline; instead, we need to subtract the baseline for the mean from all the values and divide by the baseline for the standard deviation, so we must rework our approach.

Let's write a new function, z_score_test(), to perform our Z-score tests using any number of standard deviations above the mean as a cutoff. First, we will use the get_baselines() function to calculate the baseline standard deviations by hour with the trimmed data. Then, we join the standard ...