Integrating Automation, Interactive Visualization, and Unsupervised Learning for Enhanced Diabetes Management

Effective management of diabetes necessitates efficient data handling, insightful analytics, and personalized interventions. In this study, we present a comprehensive system that automates the extraction, transformation, and loading of continuous glucose monitoring data. Data is integrated into an interactive dashboard with dual access levels: one for healthcare management professionals and another for patients for clinical management. The dashboard provides real-time updates and customizable visualization options, empowering users with actionable insights into their glucose levels. Furthermore, a clustering model to categorize patients into distinct groups based on their glucose profiles was developed. Through this model, three clusters representing different patterns of glucose control are identified. Healthcare professionals can utilize these insights to tailor treatment strategies, allocate resources effectively, and identify high-risk patients.