Akademik Veri Yönetim Sistemi
2025 Innovations in Intelligent Systems and Applications Conference, ASYU 2025, Bursa, Türkiye, 10 - 12 Eylül 2025, (Tam Metin Bildiri)
The real-time monitoring of stress levels experienced in daily life is of critical importance for individuals' health status and work productivity. The detection of stress levels through the collection of physiological data using wearable sensors is a subject addressed in the literature. However, real-time processing and interpretation of this data necessitate an infrastructure capable of handling high-volume and high-velocity. This study presents a real-time stress monitoring system based on a LightGBM machine learning model, utilizing an Apache Kafka message queue and Apache Spark stream processing infrastructure. The proposed method ingests JSON-formatted data from sensors via Kafka, processes it continuously on Spark, and generates instantaneous stress level predictions for each data segment. Data processed through Spark is written to the time-series database InfluxDB and visualized on real-time dashboards using Grafana, and through the provided interface, nurse stress levels can be monitored.