Time series often occur which exhibit some sort of "seasonality" pattern. Circadian or calendarial influences are frequently present. Derived from the generic circulant form, quasi-circulant matrices can directly and uniquely separate such seasonality from trend in the observations. The methodology is developed in APL and then applied to atmospheric lead data. Results are plotted, tabulated and discussed.
/lp/association-for-computing-machinery/exploring-time-series-analysis-using-apl-Ju5oAcKwG5