Flick the Nick: Observing Processes Over Time

Abstract

How often have you heard reference to trends in behavior over time? How often have you heard reference to the fact that something seems to be “out of control”? In many quantitative problems, the data are collected over time. In such cases, the time order of the data and the patterns gen‑erated across time are important considerations. If accidents really are on the increase, then a community might want to implement new traffic control strategies. If, however, last month’s accident total is ex­ceedingly large for some explainable reason, like road construction or bad weather, then changes in traffic control might not be warranted. For such a solution, you need to be able to explain meaningfully pat­terns in data collected over time and recognize data points that deprrt from the norm in an unusual way.