A cyclical pattern is the recurrence of fluctuations at regular or approximately regular intervals. In time-series exploration, detecting cyclicity helps distinguish recurring behavior from one-off changes. section 2.6.2
Use cyclical-pattern checks to:
- Visualize date-related variables over time before comparing business metrics across periods. section 2.6.2
- Run cyclicity detection alongside outlier checks during the first exploratory data analysis iteration. section 2.6.2
- Treat apparent fluctuations cautiously: changes may reflect operational growth, stabilization, collapse, or data-collection effects rather than a recurring cycle. section 2.6.2
Monthly review counts
Monthly review counts show minimal activity, rapid growth, and later fluctuation or collapse and rebound. These phases make time-based comparisons informative, but they do not alone establish a regular cycle. section 2.6.2
Figure 2.3 The number of reviews over time—monthly aggregation
Inspect the monthly review-count visualization to look for recurring fluctuations and distinguish them from broader trend changes. section 2.6.2
