A Scalable Data Integration and Analysis Architecture for Sensor Data of Pediatric Asthma

Area of Science:

• Biomedical research
• Data science
• Environmental health
• Pediatric health

Background:

• Asthma affects 6.8 million children in the U.S.
• Current prognostic tools are insufficient for large-scale risk investigation.
• Need for advanced data analysis infrastructure in pediatric asthma research.

Purpose of the Study:

• To present a big data integration and analysis infrastructure for pediatric asthma research.
• To enable efficient prediction and prevention of asthma attacks.
• To support biomedical researchers in investigating disease risks at scale.

Main Methods:

• Developed a big data infrastructure by the Pediatric Research using Integrated Sensor Monitoring Systems (PRISMS) Data and Software Coordination and Integration Center (DSCIC).
• Utilized Apache Kafka, Spark, Hadoop, and PostgreSQL.
• Extended Spark with a mediation layer for harmonized data analysis over heterogeneous sensor and traditional data sources, supporting both batch and stream analytics.

Main Results:

• Successfully integrated and analyzed real-time environmental, physiological, and behavioral data.
• Enabled batch and stream analytics over massive datasets from wearable and fixed sensors.
• Facilitated data analysis over a consistent, harmonized schema through logical schema mappings and query rewriting.

Conclusions:

• The PRISMS-DSCIC infrastructure addresses the limitations of current prognostic tools for pediatric asthma.
• The system empowers biomedical researchers with advanced capabilities for asthma prediction and prevention.
• Big data integration and analysis are crucial for understanding and mitigating pediatric asthma risks.