Diverse patterns of COVID-19 epidemics have been observed in populations with varied social demographics, vaccination coverages, past waves of outbreaks and public health and social measures (PHSMs) implemented in response. We discussed how nowcasting of the Omicron epidemics has been enhanced during January to August 2022 in Hong Kong by integrating evolving data streams that gradually became available and implemented the model in the nowcasting of Beijing’s Omicron BF.7 epidemics in November 2022 to January 2023. We showed that epidemic nowcasting is a useful tool if the model is parameterised with sufficient data. Our framework provides a way to integrate multiple streams of epidemiological data and knowledge for short-term assessment of the epidemic scale and disease burden.