Katkat
Back

Visualize More Than One Hundred Million Map Points

GoogleFrontend Engineer · Onsite
HardFrontend EngineerOnsiteSystem Design

Last updated

Design a multiresolution map that turns over one hundred million points into cached heat maps and detailed pins without overwhelming the browser.

Visualize More Than One Hundred Million Map Points

Design an interactive map for more than 100 million data points. Broad zoom levels should render aggregated heat-map information, while the closest zoom reveals individual pins. Discuss GIS resolution, rendering technology, zoom-aware caching, worker threads, live SSE updates, aggregation, refresh cycles, failure recovery, and hardware and network limits.

Constraints & Assumptions

  • The browser cannot download or render the entire dataset.
  • Every response is bounded by viewport, zoom, and a version or time window.
  • The design must degrade safely when live updates or high-detail data are unavailable.

Clarifying Questions to Ask

  • How fresh must the visualization be?
  • Are points static, moving, or appended over time?
  • What accuracy is required at each zoom level?

Hints

Hint 1

Make zoom a query dimension: Choose server aggregation and payload detail from viewport and zoom before comparing SVG, canvas, and WebGL.

Follow-up questions

  • How would you support historical playback?
  • What changes if some points require per-user authorization?