Hey! Just a quick question regarding the nixtla StatsForecast package. If I have a time series that does not seem to have any seasonality at all. Only some weak trend in the time series. What is the best method to use? I tried ExponentialSmoothing but it seems to only give constant predictions, I would like it to follow the trend, is this by design?

Did you try using Holt's method? There are three "versions" of exponential smoothing: • simple: only does a weighted average, so it returns a constant value • double (or Holt): considers the trend of the series • triple (holt-winters, or seasonal): considers trend and seasonality In your case, double or Holt seems to be the way to go.