Review the aligned dataset, choose the harmonization rule, and lock the transform assumptions before modeling.
Modeling frame
0 connected series staged
Harmonization: none
Transform and harmonization controls
Why use log transform?Use this when the series stays positive and moves proportionally with its level. It makes large-scale moves easier to compare with earlier ones.
When should I use a growth transform?Use a growth transform when the question is about pace rather than level. It helps compare series that live on very different scales.
Why detrend a series?Detrend when long-run drift hides the cycle you want to inspect. Leave it off if the level itself is the thing you are trying to explain.
When should I adjust seasonality?Use this when within-year seasonal patterns are obscuring the underlying signal. If the source is already adjusted, keep the current setting.
How should I handle missing values?Warn preserves the raw series, drop keeps only complete observations, and interpolate fills smaller holes. Choose the least aggressive option that still leaves you with usable data.
When do I need frequency harmonization?Use this when your inputs arrive at different frequencies. Aggregate moves faster data up to a slower cadence; interpolate smooths slower data down.
Diagnostics preview
Choose at least one connected series to open the raw-vs-analysis preview.
Readiness
Add at least one connected series or an uploaded file.
