Structural Correlation Patterns in Regional COVID-19 Surveillance Data and Implications for Epidemiological Monitoring

Abstrak

The Covid-19 pandemic has had a significant impact on the health sector in various regions, including Kabupaten Sumbawa. This study aims to analyze relationships among attributes in the Covid-19 dataset using the Correlation Matrix algorithm within the CRISP-DM methodology. The dataset was obtained from the official website of the Government of Kabupaten Sumbawa, comprising 10,573 records, of which 405 were cleaned after the data cleaning process. The analysis was conducted using RapidMiner 9.9 software. The findings indicate a very strong correlation between the attributes KONTAK ERAT-DISCARDE, SUSPEK-DISCARDE, and KONFIRMASI-MENINGGAL DUNIA with the increase in total Covid-19 cases. In addition, a significant negative correlation was observed between the attribute PP-MASIH KARANTINA and the number of deaths. Furthermore, an almost perfect correlation was found between PROBABLE-DISCARDE and PROBABLE-MENINGGAL. Based on these findings, it is recommended that the government prioritize monitoring cases before they are declared discarded and strengthen the quarantine system for travelers. This study provides a data-driven foundation for formulating evidence-based pandemic response policies.