Master class in test data schema preparation for analytics скачать в хорошем качестве

Master class in test data schema preparation for analytics

Struggling with complex SQL joins when analyzing multi-sensor data from your R&D testing equipment? In this technical webinar, Quix CTO Tomas Neubauer demonstrates two complementary schema approaches for streaming sensor data that solve the correlation problem plaguing automotive crash tests, aerospace thermal chambers, and battery testing environments. Using a live 3D printer demo with multiple sensors (temperature, fan speed, filament density), he shows how to transform MQTT topic hierarchies into both normalized and transposed formats, then sync them to cloud storage for fast analytics. Watch as we build real-time data pipelines that reduce multi-sensor query complexity from 10+ lines of complex SQL to simple 3-4 line statements, integrate seamlessly with Grafana dashboards, and enable domain experts to perform their own analysis in Jupyter notebooks. The demo covers partitioning strategies, downsampling techniques, and practical trade-offs between schema designs. Get the complete technical breakdown and implementation details in our comprehensive TL;DR guide: link https://quix.io/events/masterclass-in... 00:00 - Introduction and webinar overview 01:26 - Tomas introduction: CTO background with F1, aircraft, wind tunnel sensor data 02:30 - Previous episodes recap: 3D printer MQTT normalization 03:33 - MQTT payload problem: unusable raw format 04:20 - QuickLake database setup and S3 configuration 06:02 - Partitioning strategy: machine and sensor_id hive columns 08:49 - First deployment: normalized sensor data to S3 09:17 - Normalized schema exploration in QuickLake 10:03 - Querying normalized data structure 11:32 - Normalized schema limitations: complex multi-sensor joins 12:35 - Transposed schema introduction for better analytics 13:09 - Second pipeline setup for transposed format 17:01 - Topic configuration correction 18:01 - Transposed schema: sensors as columns 19:14 - Downsampling with time bucketing 21:06 - Advanced analytics: min/max/mean calculations 22:11 - Discoverable schema design for team collaboration 23:06 - Grafana and Jupyter notebook integration 24:14 - Live Marimo notebook update demo 25:57 - Git version control integration 26:31 - Schema approaches summary: defensive vs practical 27:33 - Data quality checks and signal validation possibilities 28:44 - Closing remarks and Q&A invitation