2 3 4 5 6 7 Patient longitudinal dat
a is delivered via the Cerner HealtheIntent platform to the HealtheDataLab data lake hosted on Amazon Simple Storage Service (Amazon S3). AWS Data Pipeline extracts the patient longitudinal data and launches an Amazon EMR cluster which processes it into metadata describing patient records, including sorting into records cohorts sharing comorbidities. Cohort data is stored as a Hive metastore on Amazon Relational Database Service (Amazon RDS). For any new patient data, Amazon CloudWatch launches an Amazon EC2 instance triggering an algorithm and ML model development workflow. Data scientists can interact with and analyze data using Jupyter notebooks. The attached EMR cluster allows model training and development. Trained and verified models and algorithms are deployed using SparkML Pipeline. Amazon EMR clusters host HealtheDataLab packages an
