11-21, 10:00–10:25 (Pacific/Auckland), WG126
Lonboard transforms geospatial analysis by eliminating the need for costly tiling and database preprocessing. This WebGL-based visualisation library delivers superior performance in Jupyter Notebooks, reducing costs while enabling real-time interaction with large vector datasets. Open source demonstration with reproducible examples from Urban Intelligence's climate modelling work.
Climate modelling generates massive raster datasets that analysts must explore to identify critical patterns like flood risk zones, temperature anomalies, and adaptation pathways. However, meaningful analysis requires overlaying large vector datasets, such as roads, infrastructure points, and building footprints, to understand the real-world impacts. Geospatial analysts traditionally face a costly bottleneck: these vector datasets must be tiled and stored in databases before meaningful visualisation can occur. This preprocessing step incurs significant computational and storage costs while creating friction in exploratory workflows.
This presentation demonstrates how Lonboard, a high-performance WebGL-based vector visualisation library, eliminates this barrier entirely, reducing workflow costs to zero while delivering superior rendering performance for the critical vector overlays. Through demonstrations using Urban Intelligence's climate analysis workflows, we'll showcase how analysts can visualise massive vector datasets directly in Jupyter Notebooks without preprocessing. Lonboard's WebGL rendering achieves speeds orders of magnitude faster than traditional packages like ipyleaflet by leveraging cutting-edge technologies, such as GeoArrow and GeoParquet, in conjunction with GPU-based map rendering.
Key benefits include elimination of costly vector preprocessing steps, instant visualisation feedback for rapid exploration, seamless Jupyter integration, and superior performance with large vector datasets essential for large datasets. The presentation features fully reproducible notebooks showcasing production-ready implementation practices.
This approach represents a fundamental shift from expensive, preprocessing-heavy workflows to cost-effective, exploration-first methodologies that accelerate workflows. By leveraging open-source tools, organisations achieve better analytical outcomes while reducing infrastructure costs and complexity.
Sam Archie is a Data Engineer at Urban Intelligence, an Aotearoa-based company that empowers communities and organisations with evidence-based tools to navigate climate change and enhance resilience. At Urban Intelligence, Sam primarily focuses on creating scalable, reliable, and efficient ETL processes using Python that power the Resilience Explorer® platform - a solution for decision-makers to navigate climate change, plan for future weather events, and enhance organisational resilience.
With a Masters of Engineering from the University of Canterbury, Sam led the implementation of software algorithms to generate strategic urban growth plans of cities in Aotearoa that minimised risk to natural hazards, increased social cohesion and reduced land prices. This unique combination of engineering background and spatial analysis expertise now enables him to build the critical data infrastructure behind intuitive geospatial mapping, visualisation of asset and network data, interactive simulations of hazard events, and risk assessments based on world-leading research.
Sam's work exemplifies how technical expertise can drive sustainable urban transformation and community resilience. His role involves transforming complex climate and hazard data into actionable insights that help communities become more resilient, healthy, and inclusive—something he never imagined during his early engineering career but now finds deeply fulfilling.