Remote sensing and social sensing for socioeconomic systems: A comparison study between nighttime lights and location-based social media at the 500 m spatial resolution

International Journal of Applied Earth Observations and Geoinformation

Yong Chen¹  Eric L. Samson²  Guofeng Cao³  Wei Zhang⁴  Naizhuo Zhao⁵ 

[1] Corresponding author at: Department of Geosciences, Center for Geospatial Technology, Texas Tech University, Lubbock, TX, 79409, USA.;Department of Geosciences, Texas Tech University, Lubbock, TX, 79409, USA;Division of Clinical Epidemiology, McGill University Health Centre, Montreal, QC, H3A 1A1, Canada;Center for Geospatial Technology, Texas Tech University, Lubbock, TX, 79409, USA;Institute of Land Resource Management, School of Humanities and Law, Northeastern University, Shenyang, Liaoning, 110169, China;

关键词: Nighttime lights imagery;    Geo-tagged tweets;    Socioeconomic factors;    Social sensing;   

DOI  :  

来源: DOAJ

【 摘 要 】

With the advent of “social sensing” in the Big Data era, location-based social media (LBSM) data are increasingly used to explore anthropogenic activities and their impacts on the environment. This study converts a typical kind of LBSM data, geo-tagged tweets, into raster images at the 500 m spatial resolution and compares them with the new generation nighttime lights (NTL) image products, the Visible Infrared Imaging Radiometer Suite (VIIRS) Day/Night Band (DNB) monthly image composites. The results show that the monthly tweet images are significantly correlated with the VIIRS-DNB images at the pixel level. The tweet images have nearly the same ability on estimating electric power consumption and better performance on assessing personal incomes and population than the NTL images. Tweeted areas (i.e. the pixels with at least one posted tweet) are closer to satellite-derived built-up/urban areas than lit areas in NTL imagery, making tweet images an alternative to delimit extents of human activities. Moreover, the monthly tweet images do not show apparent seasonal changes, and the values of tweet images are more stable across different months than VIIRS-DNB monthly image composites. This study explores the potential of LBSM data at relatively fine spatiotemporal resolutions to estimate or map socioeconomic factors as an alternative to NTL images in the United States.

【 授权许可】

Unknown