Table 1.

 Summary of forecasting and hotspot prediction tasks considered in this paper

	Forecasting	Hotspot prediction
Response variable	$Y_{\ell ,t}$ (7-d trailing average of COVID-19 case incidence rates, per location $\ell$ and time t)	$Z_{\ell ,t}=\mathbf{1}\left(Y_{\ell ,t}\ge 1.25·Y_{\ell ,t-7}\right)$ (indicator that $Y_{\ell ,t}$ grows by more than 25% relative to the preceding week)
Geographic resolution	HRR	HRR
Forecast period	9 June to 31 December 2020	16 June to 31 December 2020
Model type	Quantile regression	Logistic regression
Evaluation metric	WIS	AUC