Wearables data feed
summary
LifeOmic builds a production wearables pipeline that pulls Apple HealthKit, Fitbit, Garmin and others into a performant FHIR API using AWS Lambda, API Gateway, SQS, Kinesis and ECS batch jobs, handling duplicates and overlapping sleep records by choosing best records before ingest and using a single Kinesis shard to serialize conflicting writes. Architecture scales to ~20 million lambda invocations/month, synced 75,000 HealthKit records in proofs of concept, and uses Kinesis stream and ingestion pipeline patterns to keep latency low and costs small.
AI-written summary · about
chapters (18)
- 0:00Intro
- 0:22Meet Mark
- 1:29Mobile apps
- 2:32Weightloss journey
- 5:01Architecture
- 7:17EHR ingestion
- 9:05HTTP based EMR
- 10:31Apple Health ingestion
- 11:43OAuth
- 12:49Subscription
- 13:59Webhook payloads
- 15:58Avoiding duplicate data
- 19:24Challenges
- 23:40Overlapping records
- 27:36Architecture overview
- 30:11Lambdas
- 31:51Inversion
- 32:33Monitoring
related talks
34mModern Data Platform, IoT & the Cloud -how...
32mBuilding an IoT & Data Ecosystem for one of the Globe's Largest Container Fleets
29mUse sports data to elevate fan engagement and performance
35mStreaming and Real-time Analytics
39mLarge scale streaming analytics
38mServerless Data Streaming on AWS
42mTransform your data approach: Develop a modern data strategy
21mOptimize and manage data on Amazon S3