Compilation and Development of Big Data System for Water Management: Supporting Policy Decision-Making and Disaster Risk Management for Floods and Droughts in the Phrom Hod River Basin and Sa Kaeo Special Economic Zone

Abstract

This article aims to compile and develop a comprehensive Big Data platform for water resource management in the Phrom Hod river basin and the Sakaeo Special Economic Zone—areas identified as both environmentally vulnerable and strategically significant in Thailand’s economic and security landscape. The study integrates diverse datasets, including geographic information, climate data, water infrastructure, and socioeconomic indicators, collected from public agencies, private sectors, and real-time sensor networks. Through Extract–Transform–Load processes, the data are cleaned, harmonized, and stored in databases to enable scalable, flexible, and real-time access. The data was then used to analyze water situation trends and create linguistic models by integrating new dimensions of data from online news using high-performance web scraping techniques. This allowed for the extraction of data from 555 news websites in just 1 hour and 33 minutes through the collaboration of BeautifulSoup and Selenium, overcoming the limitations of JavaScript data loading. Furthermore, the application of the WangchanBERTa Thai language model in Named Entity Recognition (NER) demonstrated high accuracy in extracting crucial information such as date, province, and location, with an average accuracy rate of up to 85% based on human review, as well as n8n-based automation for continuous data ingestion and integration. Findings reveal that the use of Big Data significantly enhances the capacity to monitor hydrological changes, forecast flood and drought risks, and support data-driven policymaking. This integrated approach provides a robust foundation for resilient and sustainable water management at both regional and national levels, contributing to disaster mitigation and long-term water security.