Seasonality is a recurring pattern in a variable associated with particular calendar periods or cycles. In analytics, seasonal variation can create recurring daily, weekly, or monthly fluctuations that affect how metrics should be measured and interpreted. section 7.2.1 section 3.2.1
Choose an aggregation period that reduces seasonal effects without erasing meaningful variability. Larger periods can stabilize targets, but they may obscure the effects of initiatives that begin partway through the period. section 7.2.1
Use aggregation and comparison to make seasonal patterns easier to interpret:
- Inspect daily data when short-term response matters, but recognize that weekly or monthly seasonality may produce dense fluctuations. section 7.2.1 section 3.2.1
- Aggregate with sums, means, or medians when a broader time scale makes the underlying pattern clearer for stakeholders. section 3.2.1
- Compare aggregation periods with the process duration, initiative timeline, and stability of the metric before setting targets. section 7.2.1
Rat sightings in New York City
Daily 311 rat-sighting reports show an upward trend alongside consistent weekly and monthly seasonal trends. Monthly aggregation removes dense weekly spikes and highlights higher sightings during colder months, approximately November through March, versus warmer months. section 3.2.1
