Building a Better Visualization Platform with Dataprep
summary
Dataprep + Unreal Engine form a reusable pipeline that automates model cleanup, material swaps and mesh replacement, with custom Blueprint operations that do center-to-center snapping, fit corrections and randomized clutter placement. Pipelines export each configuration as JSON model-group export instead of streamed levels so the runtime spawns, manipulates and swaps kits of parts on demand. The system handles 50+ model groups, produces photoreal stills via path tracing, and supports pixel streaming for web delivery.
AI-written summary · about
chapters (11)
- 0:00Data visualization challenges
- 2:29Platform goals and needs
- 5:25Unreal Engine and Data Prep
- 7:29Data Prep filters and operations
- 9:13Blueprint operations for mesh replacement
- 14:14Configurable user interfaces
- 16:44Model groups with JSON
- 20:56Platform iteration and scalability
- 22:42Automating realism with Data Prep
- 26:38Prototype houses and engagement
- 29:11Conclusion
related talks
20m"Rebirth": Leveraging Scan Data in UE4
22mData pipelines != workflows: orchestrating data with Dagster
44mImprove your App UX with one model processing insights from batch and streams
20mBuilding a Data Pipeline with Testing in Mind
19mDemystifying Modern Data Pipeline Architecture
16mHow to Discover & Leverage Immersive DataYou Already Own
16mBeyond ETL: Revolutionizing Data Engineering Workflows
38mFrictionless Data Pipelines