Features of visit time series

This dataset contains a five-year daily time series of multiple traffic metrics collected from an educational blog on statistical forecasting, hosted under the alias statforecasting.com. The variables exhibit complex seasonal patterns driven by both weekly cycles and the academic calendar. Specifically, the dataset records daily visits over the period 2014–2020, with pronounced seasonality aligned with academic schedules in the Gregorian calendar as well as weekday effects. These characteristics make the dataset a suitable testbed for evaluating the proposed approach. The data were partitioned into 85% for training and 15% for testing.