Table 1.

Summary of results from best fit models for Tokyo, Aichi, and Osaka, 2020–2022.

	Covariate	Estimate	95% Confidence Interval
			Lower	Upper
Tokyo	Intercept	−8.676	−12.089	−5.21
	log(NP with 8-day-lag)	0.692	0.427	0.955
	Daily change of log(NP with 8-day-lag)	−2.527	−3.345	−1.713
	First order autoregression coefficient	0.968	0.95	0.986
Aichi	Intercept	−20.165	−27.325	−13.172
	log(Night Population with 9-day-lag)	1.61	1.067	2.168
	First order autoregression coefficient	0.959	0.938	0.979
Osaka	Intercept	−17.167	−28.262	−8.663
	log(NP with 8-day-lag)	1.254	0.638	2.044
	Daily change of log(NP with 8-day-lag)	−3.398	−4.92	−1.843
	First order autoregression coefficient	0.976	0.949	0.997

For Tokyo and Osaka, models with 8-day-lagged night population as well as its daily change were the best fit model, whereas the best fit model for Aichi does not include daily change in night-time population. The estimates for intercept, coefficients of explanatory variables, and the first-order autocorrelation coefficients are shown with 95% confidence intervals.