Building a movement data and analytics platform to de-stress customers & save lives
summary
A smartphone app plus a windshield IoT tag (3‑axis accelerometer and gyroscope) detects crashes and calls an emergency centre within two minutes by sending 100 Hz crash data to servers that validate the event. Ingestion uses files to S3 then Kafka running in Docker/Kubernetes, Java processors with GeoTools, storage in Postgres and reporting via Kafka → Postgres → Metabase, with TensorFlow experiments. Trial logs 330k km, ~30k trips over five months; planned rollout of 50,000 tags; challenges include many false positives from low 2g thresholds, insects, scooters and tag placement.
AI-written summary · about
chapters (10)
- 0:00Introduction and purpose
- 1:49Company context and purpose
- 3:13Problem and solution design
- 5:41Development process and principles
- 7:08Safer Journeys app and IoT device
- 9:36Data analytics platform overview
- 13:10Data collected and insights
- 14:57Applications and future use cases
- 16:25Challenges and learnings
- 18:36Q&A on precision and data usage
related talks
38mSampling and Modelling the World in Real-Time for Mobility
40mBuilding Real-Time Applications: Cyclist Crash Detection
23mMaking Sense of Streaming Sensor Data: How Uber Detects On-trip Car Crashes
50mPractical aspects of crash reporting, analytics and connectivity
27mSaving Human Lives with the IoT
19mLive Build Of A Python Service That Tracks, Transforms & Delivers Sensor Data In Real Time
40mSmart cars of tomorrow: real-time driving patterns
30mCrash detector