GlossaryStatistics and interpretation
Seasonality
Seasonality is recurring calendar-related variation. Learn how it differs from a trend and why one unusual period does not establish a pattern.
Seasonality is recurring variation associated with a repeating calendar pattern, such as time of year or day of the week. It describes a pattern across repeated cycles. One unusually busy summer or weekend is not enough by itself to establish that the pattern reliably recurs.
Look across repeated periods
A time series can rise and fall within each year while also moving upward or downward over several years. Those are different features. The recurring within-year movement may be seasonal; the longer direction may be a trend.
NIST describes seasonality as periodic fluctuation and explains that plots organized by season can help reveal recurring patterns. Its examples concern statistical analysis generally; any particular public-safety pattern still needs evidence from the relevant local data. NIST seasonality guidance.
Example
Hypothetical example: An agency's five-year series shows more park-related calls in each June through August than in each winter. The summer totals also rise gradually from year to year.
The repeated summer elevation is consistent with a seasonal pattern. The gradual rise across years is a separate possible trend. A single July-to-January comparison mixes both the season and the point in the longer series, so it cannot isolate the change over time.
Common confusion
Seasonality is not an explanation of cause. A repeating pattern can be associated with weather, use of public spaces, calendars, reporting practices, or several factors together. The chart alone does not show why the pattern occurs.
A rolling average is also not automatically a seasonal adjustment. Smoothing reduces some short-term variation, while seasonal adjustment uses a specified method to estimate and account for recurring patterns. A smooth line can retain substantial seasonality.
Make comparisons with care
Comparing corresponding months in different years can reduce some calendar differences, but holidays, leap years, and the number of weekends can still differ. For daily activity, day-of-week patterns may matter as much as month names.
Use consistent categories, boundaries, and reporting coverage across cycles. A newly published call category or a change in collection can create an apparent repeating pattern that comes from the data process. Check whether missing periods are genuinely missing rather than quiet.
If figures are seasonally adjusted, identify the method and retain the unadjusted series where useful. Do not describe expected seasonal movement as proof of improvement or deterioration without examining the longer, comparable record.