On the Benefits of Using Satellite Imagery as an Alternative Data Source in Educational Analytics

Abstract

This article examines the usefulness of using satellite images of nighttime light as alternative, accessible data in the field of education. Using images from the Visible Infrared Imaging Radiometer Suite (VIIRS), the nighttime light was measured within a 1-kilometer radius around the target schools and used to predict school construction and explain the results of students' reading functional literacy based on the international PISA Based Test for Schools (PBTS-2024) study. A random forest model with historical nighttime light data for 2012-2024 was used to predict school construction, while multilevel modeling using cross–sectional nighttime light data for 2024 was used to explain the reading literacy. The results of the study showed that the trend of nighttime light can be useful in predicting the year of school construction. At the same time, nighttime light has a statistically significant effect on the results of students' reading functional literacy. Substantively, these results show that economic activity and its dynamics around schools are important factors indicating potential locations for school construction, as well as having a positive effect on students’ educational outcomes. Thus, this study is one of the first examples of the use of alternative data, such as satellite imagery on nighttime light, in the national literature in the field of education. In practical terms, the use of nighttime light data can be used as part of the task of predicting the construction of new schools. In addition, the use of nighttime light as an additional contextual indicator of socio-economic status may be useful in developing educational policies aimed at reducing school inequality. 

Keywords: school infrastructure, satellite data, night lighting, functional literacy, forecasting, construction of new schools, multilevel modeling.