The need for early identification of regional economic conditions has substantially grown due to the diversification of the regional business cycles, primarily caused by the shifts in economic structure and the uneven impacts of the COVID-19 crisis across regions. This paper constructs the nowcasting framework, widely employed to monitor business cycles in real-time, at the regional level, with a particular focus on Busan. Given that GRDP is available only in an annual basis, we nowcast the regional economic condition index (RECI), a quarterly regional indicator developed recently by the Bank of Korea. Our RECI nowcasting model comprises a dynamic factor model as the primary sub-model and a long short-term memory model as the secondary one. The results show that the regional nowcasting models can capture shifts in the local economic condition at an early stage, typically 18 weeks prior to the RECI announcement. Furthermore, the forecasting errors are lower than those of the alternative model (autoregressive model), indicating their relatively strong predictive power. Specifically, even in the quarters of sharp economic upheaval triggered by COVID-19 shocks, our regional nowcasting models exhibit their strong predictability by capturing the shifts before the release of economic indicators. In terms of accuracy, the forecasting errors range below one percentage point throughout the forecasting horizons, further attesting to their robust predictive power.