Predicting Gross Domestic Product of Sri Lanka Using Nighttime Light Data

The main objectives of this study were to develop a model for accurately predicting Sri Lanka’s GDP using nighttime light (NTL) imagery, address existing limitations in current models, and enhance the precision of socioeconomic data extraction. The study utilised Defense Meteorological Satellite Program’s Operational Linescan System (DMSP-OLS) image data from 1992 to 2013 and Visible Infrared Imaging Radiometer Suite (VIIRS) image data from 2013 to 2019, along with Sri Lanka’s official GDP data. The total DN (Digital Number) value for each year was then extracted and correlated with Sri Lanka’s official/published GDP. The study yielded two major findings. First, while many models exist for correcting and converting DMSP and VIIRS datasets to measure socioeconomic status, some of these models fail to produce favourable results in the context of Sri Lanka. Second, the model adopted in this study measured Sri Lanka’s GDP with high accuracy. The significance of the linear regression model shows that processed data based on the model developed in the current study are more suited for estimating GDP in Sri Lanka. Therefore, it can be concluded that the NTL image processing model used in this study is a valid and accurate approach for measuring GDP using NTL data.