The role of data science is increasingly indispensable in modern economic analysis. According to Rezzy Eko Caraka, a National Research and Innovation Agency (BRIN) researcher and Research Professor at Chaoyang University of Technology, Taiwan, mastery of data science is an essential foundation for evidence-based research and public policy.
“The first step in mastering data science is to learn machine learning, which begins with mastering the basics of statistics and programming for data processing,” he explained.
At the I GET CODE 2025 Conference organised by the Association of Master’s Students in Development Economics (HIMMEP) FEB UGM on Saturday (23/8), Rezzy said that after learning machine learning, researchers can determine the focus of their studies, both in the micro and macro domains, with the support of econometrics to build and analyse models. He also emphasised that data remains at the core of analysis. Accurate and relevant data will help the government and policymakers deliver evidence-based public services while ensuring good data management practices.
Rezzy mentioned that data can be obtained from various channels, such as mobile and social media data, academic publications, financial data from Bloomberg or other institutions, government data, and multimedia data. Data can also be collected independently through surveys, experiments, laboratory results, and satellite data.
One example of data he highlighted was satellite data obtained through remote sensing, or techniques for collecting geospatial data through satellites or aerial vehicles such as drones, balloons, or aircraft. This technology produces big data with massive and multidimensional coverage, which helps analyse infrastructure, urban footprints, and agricultural areas. This data can be accessed through international institutions such as NASA via the Earth Science Data System (ESDS).
Rezzy explained that big data is broad, complex, and sourced from various channels. The main challenge is to maintain the value of the data despite its general nature, then process it with technology, mathematics, and computation-based analytics. This process allows researchers to discover hidden patterns, implicit relationships, and essential information supporting decision-making. Data science enables the quantification of human behaviour, providing insights into market segmentation, consumer behaviour, and marketing strategies.
Rezzy presented one of his studies on using AI for early cancer detection and planning cancer centres in Indonesia. The study used remote sensing data to analyse factors such as elevation, slope, temperature, distance, and proximity to water sources to identify the most suitable locations for cancer facilities. In addition, big data analysis also enables the estimation of a population’s vulnerability to cancer in a given area, which is a basis for consideration in the development of health service centres.
However, Rezzy reminded us that AI should not be used as the sole determinant of policy but rather as a research support tool. This is because decisions must remain in the hands of humans, as the validation of data accuracy and logic cannot be completely replaced by machines.
“AI systems should be designed to provide indications, patterns, or anomalies. However, final decisions must still be made by humans, including through manual cross-checking processes,” he concluded.
Reportage by: Najwah Ariella Puteri
Editor: Kurnia Ekaptiningrum
Sustainable Development Goals
