Quantitative spatiotemporal impact of dynamic population density changes on the COVID-19 pandemic in China’s mainland

Figures & data

Figure 1. Tencent Population Density and estimated population density in 2015. (a) is the Tencent positioning points at 12 A.M. on 1 December 2019. (b) shows an estimated population density distribution in 2015. (c) shows the 32-levels’ classified results of the Average Tencent Population Density (ATPD) during the study period. (d) are the 32-levels’ classified density distribution corresponding to (b).

Table 1. Variable definition, source, and statistic description

Domain	Variable	Description	Source	Mean	Std.	Obs.
Dependent variable
Confirmed Case	NC	New confirmed case number of the day	Bao et al. (Bao, Lin, and Zhou 2020) gathered from the National, Provincial and Autonomous Region’s Health Commission	1.54	67.66	52,780
	NC_Lag5	New confirmed case number lagging 5 days		1.54	67.66	52,780
	NC_Lag14	New confirmed case number lagging 14 days		1.54	67.66	52,780
Death Case	ND	New death case number		0.06	2.00	52,780
	ND_Lag5	New death case number lagging 5 days		0.06	2.00	52,780
	ND_Lag14	New death case number lagging 14 days		0.06	2.00	52,780
Independent variable
People Activity	PD	Tencent population density	Tencent Positioning Data Platform (2020)	$2.83\times {10}^{54}$	$4.59\times {10}^{56}$	52,780
	NPD	New Tencent population density, $\mathit{NPD}=ln\left(\mathit{PD}+1\right)$		54.39	20.54	52,780
Socioeconomic	BN	Hospital beds number per thousand people	China Health Statistics Yearbook 2020 (2021)	6.32	0.83	52,780
	AO	The number of people whose age are over 65		11.19	2.50	52,780
Environmental (Air Quality)	PM25	Average PM2.5 concentration (μg/m^3) in the day.	National Urban Air Quality Real Time Release Platform (2021)	40.22	32.67	52,780
	PM10	Average PM10 concentration (μg/m^3) in the day.		64.00	61.48	52,780
	SO2	Average SO2 concentration (μg/m^3) in the day.		10.13	7.07	52,780
	CO	Average CO concentration (μg/m^3) in the day.		0.76	0.37	52,780
	NO2	Average NO2 concentration (μg/m^3) in the day.		23.46	13.59	52,780
	O3	Average O3 concentration (μg/m^3) in the day.		47.86	20.53	52,780
Environmental (Weather Condition)	TP	Average temperature (°C) in the day.	NOAA National Centers for Environmental Information (2020)	7.71	8.67	52,780
	DP	Average dew point (°C) in the day.		1.97	8.35	52,780

Note: Std. refers to standard deviation. Obs. refers to the number of observations.

Figure 2. Simple linear regression between ATPD and EPD.

Table 2. Monthly global Moran’s I for dependent variables

	NC	NC_Lag5	NC_Lag14	ND	ND_Lag5	ND_Lag14
Dec. 2019	.	.	.	.	.	.
Jan. 2020	0.130	0.099	0.045	0.035	0.037	0.028
Feb. 2020	0.031	0.021	0.010	0.024	0.023	0.023
Mar. 2020	.	.	.	0.022	0.014	0.008
Apr. 2020	0.129	0.168	0.180	.	.	.

Note: The insignificant values have been shown as “.”.

Table 3. SDM pre-robustness test and effect selection

Spatial Matrix	Dependent Variable	SAR p-value	SEM p-value	Hau. p-value
Contiguity	NC	< 0.01	< 0.01	0.66
	NC_Lag5	< 0.01	< 0.01	0.74
	NC_Lag14	< 0.01	< 0.01	0.66
	ND	< 0.01	< 0.01	0.56
	ND_Lag5	< 0.01	< 0.01	0.67
	ND_Lag14	< 0.01	< 0.01	0.32
Inverse Distance	NC	< 0.01	< 0.01	0.73
	NC_Lag5	< 0.01	< 0.01	0.80
	NC_Lag14	< 0.01	< 0.01	0.79
	ND	< 0.01	< 0.01	0.93
	ND_Lag5	< 0.01	< 0.01	0.89
	ND_Lag14	< 0.01	< 0.01	0.55

Note: Hau. refers to Hausman test.

Figure 3. Main coefficients of NPD. (* p < 0.05, ** p < 0.01, *** p < 0.001).

Table 4. Direct and indirect effects of variables on COVID-19 transmission

Indep. Var.	Eff.	Contiguity Matrix						Inverse Distance Matrix
		NC	NC_Lag5	NC_Lag14	ND	ND_Lag5	ND_Lag14	NC	NC_Lag5	NC_Lag14	ND	ND_Lag5	ND_Lag14
NPD	Dir.	0.101*	0.131**	0.154**	0.00366	0.00423*	0.00749***	0.0944*	0.118*	0.138**	0.00413*	0.00444*	0.00693***
	Ind.	−0.207**	−0.160*	−0.205**	−0.00787*	−0.00879**	−0.00739*	−0.712**	−0.337	−0.566*	−0.0212**	−0.0200**	−0.011
BN	Dir.	−0.235	−0.264	0.347	0.00123	−0.0101	0.0206	1.913	1.647	2.272	0.0884	0.0823	0.1
	Ind.	0.638	1.156	0.365	0.0212	0.0354	0.0285	−0.998	3.161	0.164	0.0575	0.0876	0.137
AO	Dir.	−0.356	−0.391	−0.472	−0.0118	−0.0146	−0.0223	−0.967	−0.734	−0.985	−0.0365	−0.0364	−0.0375
	Ind.	1.021	0.786	0.848	0.0357	0.0409	0.0296	3.701	1.715	2.388	0.0732	0.0748	0.00334
PM25	Dir.	0.0104	−0.00583	−0.0167	0.000307	−0.00009	−0.00077	−0.0154	−0.0465*	−0.0612**	−0.0003	−0.00111	−0.0025***
	Ind.	−0.0297	0.0686	0.0691	−0.00279*	−0.00137	0.000575	−0.0911	0.206	0.439***	−0.00843*	−0.00402	0.00942**
PM10	Dir.	−0.00296	8.03 $\times$ 10^−5	0.00335	−2 $\times$ 10^−7	6.92 $\times$ 10^−5	0.000196	0.00724	0.015	0.0186	0.000178	0.000437	0.000812**
	Ind.	−0.00393	−0.0444*	−0.0333	0.000908	0.000233	−0.00076	−0.0559	−0.166**	−0.189**	0.00340*	3.48 $\times$ 10^−5	−0.00544***
SO2	Dir.	0.0404	0.00786	−0.0527	0.00246	0.00167	−0.0003	−0.0564	−0.0953	−0.0946	−0.00133	−0.00265	−0.00394*
	Ind.	0.168	0.256	0.146	−0.00234	0.00234	0.00571	1.522***	1.384**	0.299	0.0468***	0.0537***	0.0377**
CO	Dir.	−0.156	1.916	2.531	−0.0338	0.012	0.0829	−2.113	−0.375	0.552	−0.0798	−0.0539	0.00599
	Ind.	7.607*	0.65	4.002	0.338***	0.222*	0.19	15.55	2.162	−2.616	0.548*	0.414	0.0532
NO2	Dir.	−0.0189	−0.0214	−0.00997	−0.00038	−0.00037	−0.00038	0.148***	0.155***	0.113**	0.00568***	0.00673***	0.00633***
	Ind.	−0.287***	−0.270***	−0.216***	−0.00744***	−0.00894***	−0.00842***	−0.717***	−0.630***	−0.339**	−0.0278***	−0.0287***	−0.0209***
O3	Dir.	0.0242	0.0208	0.0212	0.000358	0.000798	0.000783	0.0262	0.0245	0.0183	0.000545	0.00117*	0.0009
	Ind.	−0.0297	−0.0279	−0.0167	0.000501	−0.00084	−0.00024	0.0047	0.0385	0.0661	−0.00187	−0.00206	0.00146
TP	Dir.	−0.0609	−0.0738	−0.0577	−0.00105	−0.00084	−0.00052	−0.0479	−0.0678	−0.0408	−0.00079	−0.00057	−0.00024
	Ind.	0.159	−0.00123	0.573*	0.00442	0.00679	0.0152*	2.828	−1.053	3.277	0.126	0.0846	0.0755
DP	Dir.	0.0355	0.0407	0.037	0.000289	0.000323	7.62 $\times$ 10^−6	0.0277	0.0409	0.0222	0.00015	0.000189	−0.0002
	Ind.	−0.182	0.108	−0.580**	−0.0037	−0.00477	−0.0127*	−2.247	1.191	−3.234	−0.112*	−0.0623	−0.0718

Note: Indep. Var. refers to the Independent variable. Eff. Dir. and Ind. refers to the direct and indirect effects. * p <0.05, ** p <0.01, *** p <0.001.

Figure 4. The relationship between EDPD and COVID-19.

Table 5. $\mathit{R}{\mathit{I}}_{\mathit{coef}.}$ and $\mathit{R}{\mathit{I}}_{\mathit{sig}.}$ of VS1.

		Contiguity Matrix								Inverse Distance Matrix
	NC		NC_Lag5	NC_Lag14	ND	ND_Lag5	ND_Lag14		NC	NC_Lag5	NC_Lag14	ND	ND_Lag5	ND_Lag14
Main
$\mathit{R}{I}_{\mathit{coef}.}$	61.50%		61.68%	65.35%	94.16%	95.19%	83.82%		66.00%	64.30%	66.53%	97.60%	93.58%	88.31%
$\mathit{R}{I}_{\mathit{sig}.}$	40%		100%	100%	20%	20%	100%		20%	100%	100%	20%	20%	100%
Dir.
$\mathit{R}{I}_{\mathit{coef}.}$	62.59%		61.79%	65.06%	92.65%	94.02%	83.75%		65.53%	64.46%	67.39%	97.68%	93.81%	88.41%
$\mathit{R}{I}_{\mathit{sig}.}$	40%		100%	100%	20%	20%	100%		10%	100%	100%	20%	20%	100%
Ind.
$\mathit{R}{I}_{\mathit{coef}.}$	55.95%		60.06%	62.66%	58.35%	63.36%	73.86%		61.85%	62.29%	31.42%	83.63%	88.80%	66.96%
$\mathit{R}{I}_{\mathit{sig}.}$	100%		20%	100%	70%	100%	50%		80%	10%	90%	50%	50%	20%

Note: Dir. and Ind. refers to the direct and indirect effects.

Table 6. SDE model results of VS2

	Contiguity Matrix				Inverse Distance Matrix
	NC	NC_Lag5	ND	ND_Lag5	NC	NC_Lag5	ND	ND_Lag5
Main	2.674***	1.689*	0.137***	0.102***	3.166***	2.114**	0.164***	0.121***
Cons.	+/+	+/+	+/+	+/+	+/+	+/+	+/+	+/+
Dir.	2.628***	1.659*	0.134***	0.0997***	3.123***	2.083**	0.161***	0.119***
Cons.	+/+	+/+	+/+	+/+	+/+	+/+	+/+	+/+
Ind.	−3.605*	−2.467	−0.183***	−0.139***	−30.84***	−23.22**	−1.687***	−1.201***
Cons.	+/+	+/-	+/+	+/+	+/+	+/-	+/+	+/+

Note: Dir. and Ind. refers to the direct and indirect effects. Cons. indicates the consistency. The left sign of the consistencies represents the consistency of positive/negative, while the right one represents the consistency of significance. “+” means they are consistent, while “-” means they are contrary. For example, “Cons.” under “Ind.” on the second column is “+/+”, which means the indirect effects coefficient of BMNI and NPD for NC_Lag5-SDE models are consistent (“+”), they are all positive, while the right one for the significance are inconsistent (“-”), the BMNI one is insignificant, while the NPD one is significant.

Bao, J., H. Lin, and Q. Zhou. 2020. “COVID-19 Epidemic Spatio-Temporal Dataset User Guide.” Accessed Feb 18. https://github.com/Estelle0217/COVID-19-Epidemic-Dataset.git .

 Google Scholar

Tencent “Tencent Positioning Big Data.” https://heat.qq.com/ . Accessed 1 June 2021.

 Google Scholar

National Health Commission of the People’s Republic of China. “China Health Statistics Yearbook 2020 .” http://www.nhc.gov.cn/zwgk/tjnj1/ejlist.shtml. Accessed 1 June 2021.

 Google Scholar

National Urban Air Quality Real Time Release Platform http://www.cnemc.cn/sssj/ . Accessed 1 June 2021.

 Google Scholar

NOAA National Centers for Environmental Information https://www.ncdc.noaa.gov/ . Accessed 1 June 2021.

 Google Scholar

Data Availability Statement

The authors confirm that the data and the code supporting the findings of this study are available within the article and its supplementary materials.