New Zealand Statistical Association 2024 Conference

Jamas Enright

Stats NZ

Synthetic bridge over troubled data

The recent COVID pandemic disrupted a lot of time series. For New Zealand, the impact was limited to either one period or just a few years. With the series returning to pre-COVID levels, we needed a way to side-step the COVID impacted data, to maintain the seasonal patterns from pre-COVID to the post-COVID period. This talk will explain how we create a synthetic bridge that we use in our seasonal decomposition system to replace the disrupted data with data to maintain that seasonal pattern.