Seasonality refers to fluctuations or variations in a time series that occur periodically due to factors such as weather, holidays, or economic cycles. In the context of mobile games this could include strong annual seasonality patterns driven by players being more active in summer, or weekday seasonality driven by weekly events or players being more active at the weekend. The Kohort platform uses Machine Learning (ML) to detect seasonality patterns within the historical data and apply these to the forecast, as well as to de-season the model input to ensure seasonal effects in recent data aren’t mistaken for an upward or downward trend.