The relationship between nighttime light intensity and GDP in Shanghai districts

Abstract

Night light data can well reflect the economic prosperity and exploring the relationship between the nighttime light intensity and GDP can have a better understanding on the impact of night economy on regional GDP. After rectifying two satellites’ data, regression analysis was used to study the relationship between the average intensity of nighttime lights and GDP. The consequence showed that the GDP of Pudong New District, Chongming District and Shanghai is the most closely related to the average intensity of night light and presenting corresponding advice. It is the first application of multiple math models and calculations within the research field of nitghttime light intensify and GDP in Shanghai, which is a try in practice that has gained well calculated results and performed creativities.

Keywords

Shanghai night light intensity GDP

1. Introduction

Shanghai is the central of economy, finance, trade, shipping and technology of China and it is also one of the most flourishing cities in China. The night scene is charming as one of the populist tourisms. The essence of night light is the lights and fires generated by production, business, and life at night, so the degree of night light reflects the economic activity of Shanghai. Recently, most researches focused on the realm of the whole nation and provincial scope, while research on the relationship between nighttime light data and GDP in different areas of a city is helpful to formulate urban plannings and policies, especially development policies of different administrative boundaries, thus, it has good practical significance.

2. Literature review

There are more studies on the relationship between nighttime light data and GDP in academia, with different research methods and ideas. Jia and Qin [1] constructed GDP prediction model to estimate the GDP value of Beihai, Fangcheng port and Qinzhou, and analyzed the spatial distribution characteristics of GDP, which based on the DMSP/OLS nighttime light data.

Yuan and Chen [2] came up with a new method that based on analyzing the relationship between nighttime light intensity and GDP, a spatial model of GDP was established by using the third-order fitting numerical analysis method. The results showed that the accuracy error of third-order numerical fitting is not more than 3.9%. The analysis efficiency of total province GDP is high, through the nighttime light data measurement. It can verify the authenticity of the economic growth and GDP statistics of inner Mongolia province from a relatively objective perspective. Based on NPP-VIIRS nighttime light data products, the correlation between comprehensive light index CNLI and GDP of Chengdu is obtained through correlation and regression analysis, and the regression model between CNLI and GDP is established [3]. After correcting the preliminary simulated GDP values of each pixel through linear adjustment, the spatial density map of GDP in Chengdu from 2012 to 2018 was born. Through error analysis, it was found that the relative error between the simulated GDP and the statistical GDP of each administrative region was subtle, and the overall accuracy was high. The results showed that the high-value area of Chengdu’s GDP appeared at the junction of Qingyang District and Jinjiang District, and gradually decreased radially in all directions. Meanwhile, during the research period, the areas with high GDP growth are mainly concentrated in the downtown area of Chengdu. Chai et al. [4] compared the adaptability of DMSP-OLS and NPP-VIRS nighttime light data in small-scale unit GDP estimation and confirmed that NPP-VIIRS data had better performance in township-level GDP estimation. The revised NPP-VIIRS nighttime light data was used to establish the estimation model of township-level GDP in the Pearl River Delta region, and the estimated results were corrected by regional difference coefficient. The overall accuracy of the experimental results of township-level GDP estimation in 2013 reached 85%. Li et al. [5] firstly removed isolated extremely bright pixel and background noise. Then they made the correlation analysis of total light intensity, linear weighted light index, and comprehensive light index with GDP of 11 districts in Hebei Province, which was based on the NPP-VIIRS nighttime light intensity. The consequence presented the correlation between the total light intensity at night and the GDP of each district in the city is the most significant. The GDP corresponding to each pixel was calculated according to the regression model corresponding to the highest correlation coefficient in each city. After linear correction, the GDP density map of Hebei Province was generated. Li et al. [6] conducted linear regression analysis on them with SPSS software through DMSP/OLS nighttime light data and GDP data of Guangxi in 2005, 2008, 2010 and 2013, and the results showed that nighttime light data of these four years had an obvious linear relationship with GDP in both provincial and municipal areas. Wu [7] took advantage of DMSP/OLS nighttime light data of 2003, 2005, 2007, 2009, 2011 and 2013 in China and GDP of corresponding years to discuss the relationship between nighttime light and GDP in Chinese cities from 2003 to 2013. The results showed that the nighttime light data from 2003 to 2013 had an obvious relationship with GDP in both the national and provincial scales. Chen et al. [8] conducted correlation analysis on DMSP/OLS nighttime lighting data and Henan Province’s GDP, and obtained the spatial model. And they verified the accuracy of this, and generated the GDP density map of Henan province and the map of GDP density growth in both Henan Province and Zhengzhou City. By using NPP-VIIRS nighttime light data, which is more accurate than DMSP-OLS light data, Yong and Liu [9] analyzed the authenticity of GDP data of prefecture-level cities in China from a more objective perspective. From the subjective level, the scatter diagram of urban lighting data and GDP in east China and the comparison diagram of lighting data in 1995 and 2013 were analyzed, and it was concluded that there was a high correlation between GDP and lighting data. Based on the characteristics of night lighting GDP statistics data and comparative analysis [10], the system combed the economics based on night light data adjustment and correction of GDP statistics literature. Scholars domestic and abroad were reviewed in this paper about how to use the night light data gap on economic growth and studied the theory of spatial distribution of economic activity.

3. Data sources and processing

3.1 Nighttime light data

Nighttime light data are divided into DMSP-OLS and NPP-VIIRS data. However, the two remote sensing data are not comparable. Hence, the two sets of data cannot be used at the same time. DMSP-OLS and NPP-VIIRS, these two generations of mainstream nighttime light data have different resolutions, spectral response modes, and other parameter information. The mismatch of parameter information leads to the occurrence of time sequence fault of nighttime light data [11]. Therefore, DMSP-OLS and NPP-VIIRS data were corrected and processed in a unified manner to construct a multi-source night light remote sensing dataset of Shanghai with long time series from 2001 to 2020 in this paper. Specific as follows:

The average nighttime light intensity (light density) of a region is a good indicator of the lighting characteristics of the region. In this paper, the average light intensity (light density) is selected to represent the nighttime light data of each district in Shanghai. The Total Nighttime Light Index (TNLI) was constructed to calculate the Average Nighttime Light Intensity (ANLI) in the region, and formulas as follows:

$\displaystyle\textit{TNLI}=\mathop{\sum}\limits_{i=1}^{n}\textit{DNi}$ (1) $\displaystyle\textit{ANLI}=\frac{\textit{TNLI}}{n}$ (2)

In the formula:

DNi is the pixel radiation value of each grid cell in the region; N is the number of grids in the region; TNLI is the total amount of light in the area; ANLI is the average light intensity in a certain area.

3.2 GDP of Shanghai districts

The data of each district in Shanghai in this paper are from 2012 to 2021“Shanghai Statistical Yearbook”, “Shanghai Pudong New Area statistical yearbook”, “Shanghai Huangpu District statistical yearbook”, “Shanghai Minhang District statistical yearbook”, “Shanghai Jiading District statistical yearbook”, “Shanghai Jing’an District statistical yearbook”, “Shanghai Xuhui District statistical yearbook”, “Shanghai Yangpu District statistical yearbook”, “Shanghai Songjiang District statistical yearbook”, “Shanghai Baoshan District statistical yearbook”, “Shanghai Qingpu District statistical yearbook”, “Shanghai Fengxian District statistical yearbook”, “Shanghai Putuo District statistical yearbook”, “Shanghai Changning District statistical yearbook”, “Shanghai Hongkou District statistical yearbook”, “Shanghai Jinshan District statistical yearbook”, “Shanghai Chongming District statistical yearbook”, district statistical bulletins and census reports. The statistics data are shown in the following Table 1.

4. Regression analysis of the relationship between nighttime light intensity and GDP in Shanghai districts

To explore the relationship between night light intensity and GDP in 16 districts of Shanghai, this paper selects the regional GDP from 2011 to 2020 and the average night light intensity I for regression analysis. The regression model is expressed as follows:

Table 1
GDP of 16 districts of Shanghai from 2011 to 2020 (Unit: 100 million Yuan)

Year	Pu dong new district		Huangpu district		Minhang district		Jiading district		Jing’an district		Xu hui district		Yangpu district		Songjiang district		Baoshan district		Qingpu district		Fengxian district		Pu tuo district		Changning district		Hongkou district		Jinshan district		Chongming district
2011	5484	.35	1291	.57	1483	.08	914	.8	582	.34	956	.67	1026	.89	934	.17	745	.73	665	.2	572	.6	619	.98	720	.74	610	.39	425	.74	224	.1
2012	5929	.91	1368	.42	1594	.22	969	.7	618	.67	1006	.59	1201	.11	886	.55	824	.2	732	.31	624	.6	657	.58	752	.73	659	.22	553	.5	236	.3
2013	6448	.68	1475		17221		1050	.6	665	.27	1095		1350		917	.49	897	.5	771	.9	644	.8	721	.92	795		720		659	.22	252	.3
2014	7109	.74	1624	.74	1843	.75	1630	.2	732	.07	1187		1467	.86	969	.46	964	.61	827	.4	668	.4	787	.74	929	.18	762		726	.5	272	.2
2015	7898	.35	1700		1964	.71	1756	.1	1480		1300		1581	.47	995	.36	1000	.6	878	.2	685	.8	838	.74	1046	.58	800		867	.02	291	.22
2016	8732	.5	2019	.2	2101	.2	1875	.9	1662	.2	1522	.2	1629	.5	1040	.4	1048	.7	939	.7	727	.2	973	.9	1316	.2	889	.8	922	.9	311	.7
2017	9651	.39	2104	.44	2237	.39	2151	.7	1703	.68	1573	.94	1703	.19	1120	.97	1147	.36	1009	.23	779	.33	933	.46	1317	.52	780	.49	978	.32	332	.84
2018	10460	.09	2270	.31	2013	.65	2362	.7	1847		1667	.39	1847	.76	1279	.67	1392	.06	1074	.3	841	.5	1001	.77	1418	.78	838	.01	759	.87	351	.1
2019	12734	.2	2577	.9	2520	.8	2608	.1	2298	.7	2111		2083	.2	1579	.7	1551	.5	1166	.3	1173	.2	1111	.6	1649	.1	1033		1077	.2	378	.5
2020	13207	.03	2650		2565		2488	.4	2323	.08	2176	.73	2106	.63	1637	.11	1578	.48	1194	.01	1190	.19	1129	.51	1097	.36	1047	.28	828	.6	381	.83

$\displaystyle\textit{GDP}_{i}=P+aI$ (3)

$P$ and $a$ are the coefficient of regression model, $I$ is the average intensity of light at night.

In this paper, the correlation coefficients of the city and 16 districts in Shanghai with the average intensity of night light are calculated successively, and the regression fitting equation I between the GDP of the whole city and 16 districts and the average intensity of night light is established, see Table 2.

Table 2

Regression fitting results of GDP of Shanghai and 16 districts with average intensity of night light I

Number	District	Regression model	R		Fitting precision
1	Pudong	GDP $=$ 508.01I $-$ 578.24		0.9079929	0.8245
2	Huangpu	GDP $=$ 69.346I $-$ 1837.4		0.5296318	0.2805
3	Minhang	GDP $=$ 207.32I $-$ 1849.5		0.2312419	0.0535
4	Jiading	GDP $=$ 139.75I $-$ 879.03		0.6796938	0.4620
5	Jing’an	GDP $=$ 113.5I $-$ 3112.7		0.6375558	0.4065
6	Xuhui	GDP $=$ 66.784I $-$ 937.6		0.5452691	0.2973
7	Yangpu	GDP $=$ $-$ 41.285I $+$ 3003.8	$-$	0.316071	0.0999
8	Songjiang	GDP $=$ 56.667I $+$ 379.45		0.5953684	0.3545
9	Baoshan	GDP $=$ 60.5I $-$ 34.104		0.7837826	0.6143
10	Qingpu	GDP $=$ 54.162I $+$ 350.46		0.8108193	0.6574
11	Fengxian	GDP $=$ 72.405I $+$ 209.95		0.830761	0.6902
12	Putuo	GDP $=$ 26.028I $-$ 96.863		0.7592877	0.5765
13	Changning	GDP $=$ 23.842I $-$ 7.9834		0.6877898	0.4731
14	Hongkou	GDP $=$ $-$ 10.763I $+$ 1218.5	$-$	0.249117	0.0621
15	Jinshan	GDP $=$ 74.28I $+$ 246.94		0.7521139	0.5657
16	Chongming	GDP $=$ 48.364I $+$ 232.98		0.8641363	0.7467
17	Shanghai	GDP $=$ 2016.5I $+$ 4121.5		0.8719274	0.7603

Unit: 100 million Yuan.

Figure 1.

Regression fitting results of Shanghai GDP modeling.

As it shown in Table 2, R ${}^{2}$ of Pudong New Area, Chongming District and Shanghai are high, which are above 0.700 while the other 14 districts are low. The fitting accuracy (R ${}^{2}$ ) of Minhang District, Yangpu District and Hongkou District was lower than 0.1. Thus, the GDP of Pudong New Area, Chongming District and Shanghai is most closely related to the average intensity of night light. The above results indicate night production, operation and living activities in Pudong New Area and Chongming District contribute more to GDP, while the daytime economy in Minhang District. Yangpu District and Hongkou District are more developed, and the impact of night economy on GDP is relatively less. In general, as far as Shanghai is concerned, there is a close correlation between GDP and night light intensity, implicating that the overall night economy of Shanghai is relatively developed and contributes greatly to GDP. Figure 1 shows the regression fitting results of GDP and night light intensity in Shanghai.

5. Conclusion

It is widely known that since 1930s, the reputation of “Nightlife in Shanghai” has swept around the world. Shanghai is a fusion of modern and tradition with the accumulation of historical cultures, international marks, and diverse collections. Nearly a century later, night Shanghai has been given a new definition in the new era, and the “night economy” has been evaluated as a trillion-level market opportunity. Shanghai still lags other Chinese cities such as Shenzhen and Chengdu in indicators of its nighttime economy, such as the number of bars and late-night movies. Therefore, Shanghai should further promote the development of night economy, so that enhance the night economy and enrich the night business. For instance, night safari park, night museum, night bookstore and other night businesses. Also, attention should be paid to policies and measures in traffic and urban management to optimize the quality of night economy.

References

Jia

Qin

. Analysis of GDP forecast in Guangxi coastal area based on DMSP/OLS nighttime light data. Urban Constr Guangxi. 2021; (1): 8-12.

Yuan

Chen

. Gross GDP analysis of Inner Mongolia Province based on night light data. J Anyang Norm Univ. 2016; (5): 83-86.

Bai

Tang

. Spatialization of GDP in Chengdu based on NPP-VIIRS nighttime light data. Surv Mapp. 2020; 43(4): 147-151.

Chai

Wang

Qiao

. Estimation of township-level GDP in the Pearl River Delta region based on night light data. Trop Geogr. 2015; 35(3): 379-385.

Wei

Sun

. Spatialization of GDP in Hebei Province based on NPP-VIIRS nighttime light data. J Xinyang Norm Univ: Nat Sci Ed. 2016; 29(1): 152-156.

Wei

Tong

. Analysis on the relationship between night light and GDP – Taking Guangxi Zhuang Autonomous Region as an example. Public Sci Technol. 2020; 22(12): 140-143.

. Discussion on the relationship between nighttime light and GDP in Chinese cities from 2003 to 2013. Housing Real Estate. 2018; 520(34): 243-245.

Chen

Wang

. Spatialization of GDP in Henan Province based on night light data. Areal Res Dev. 2020; 39(1): 28-31+45.

Yong

Liu

. The authenticity of GDP data of Prefecture-level cities in China: A test based on NPP-VIIRS Nighttime light Data. Stat Decis. 2019; 35(13): 19-23.

10.

Wang

Huang

. Night light data and its application in economic research. Econ Trends. 2018; (10): 75-87.

11.

Yang

. Research on urban agglomeration development process based on DMSP/OLS and NPP/VIIRS Integrated Data. Southwest University; 2018.

The relationship between nighttime light intensity and GDP in Shanghai districts

Abstract

Keywords

1. Introduction

2. Literature review

3. Data sources and processing

3.1 Nighttime light data

4. Regression analysis of the relationship between nighttime light intensity and GDP in Shanghai districts

Table 1 GDP of 16 districts of Shanghai from 2011 to 2020 (Unit: 100 million Yuan)

References

Table 1
GDP of 16 districts of Shanghai from 2011 to 2020 (Unit: 100 million Yuan)