Mapping Gridded GDP Distribution of China Based on Remote Sensing Data and Machine Learning Methods

Abstract

The gridded spatial distribution data of Gross Domestic Product (GDP) has a wide range of application values in many fields, such as regional economic analysis, urban planning, sustainable utilization of resources, and disaster risk assessment. However, currently the publicly accessible GDP grid datasets face limitations in terms of temporal coverage, spatial extent, and accuracy. Therefore, based on the remote sensing data of land use and nighttime light, this study developed two methods: the factor averaging method (FAM) and grid averaging method (GAM), and used Random Forest (RF) and eXtreme Gradient Boosting (XGBoost) algorithms to jointly construct the spatial model of GDP, so as to produce China’s 1 km gridded GDP in 2020. The experimental results show the following: (1) The GAM yields higher R2 values than the FAM in modeling the three industries, and therefore, it is adopted as the basis for GDP spatialization modeling. (2) XGBoost achieves higher R2 values than RF in modeling primary and secondary industries, but lower R2 values in modeling tertiary industry. Consequently, both methods are combined to construct the overall GDP spatialization model. (3) The accuracy of the GDP spatialization results is evaluated based on town-level GDP statistics, with an R2 value of 0.78, indicating its reliable predictive capability. (4) Compared with publicly available GDP datasets, our dataset exhibits consistent spatial distribution patterns and aggregation trends. Furthermore, our GDP dataset provides a more detailed depiction of variations within county-level administrative units. Therefore, the method proposed in this study offers a valuable option for generating a gridded GDP dataset, visually displaying the uneven economic development across various regions in China. It helps to uncover economic disparities among regions and provides data support for formulating differentiated support policies, so as to promote balanced regional development among regions. Furthermore, it contributes to promoting sustained, inclusive, and sustainable economic growth (SDG 8) and reducing inequalities within and among countries (SDG 10), thereby providing strong support for urban planning and sustainable development.