Introduction

From my previous topic about Building Urban Digital Twins – From a Building-level to National-level, I mentioned Unity as our primary engine for developing our digital twin projects.

What we did from 2021

We scored Unity, Unreal, WebGL, Omniverse, Vizzio, 51World and several others against a weighted matrix. Unity topped the list because it balanced real-time graphics, a vast plugin ecosystem, and cross-device build targets. Vizzio and 51World were evaluated only on paper and were never used in production.

Since, the primary goal of the application is to achieve both very good visuals and usability within enterprises. Unity allows us to achieve both points, and at the same time, it allows us to make the application portable as we can run the application on both desktop and browser (via WebGL).

However, a known caveat is that there are certain performance limitations for us when it comes to rendering. Post-processing effects such as bloom, depth-of-field and motion-blur pipelines are getting stripped so the project falls back to low-quality paths, worse case, crash. So, we have to keep a separate branch for a lightweight, WebGL version.

Criteria we used on selecting platform

There are several other criteria aside from the above factors which may vary from case-to-case.

What changed by 2025

• WebGPU lifts browser visuals far above WebGL (Vagon)
• Unity Web build now targets WebGPU, reducing the need for local installs. (Unity Docs)
• Unreal Pixel Streaming adds AV1 support-same visuals, lower bitrate. (Epic Docs)
• Cesium plug-in streams real-world 3D tiles into Unity. (Cesium)

Decision reference matrix

5. What hasn't changed much

Even with these platforms available, the readiness to consume and visualize data hasn’t changed much-each project still needs its own custom development work.

Libraries exist for data-viz (e.g., deck.gl on the web, plus counterparts for Unity, Unreal, Omniverse, etc.), but every platform usually needs a different library for each visualization type.

That’s why I believe open schemas like Cesium’s 3D Tiles and Omniverse’s USD are the future for true interoperability.

Soon, non-technical users should be able to present geospatial data as easily as creating a PowerPoint slide.

6. How I decide now

1. Rank the criteria by project needs.
2. Plug the weights into this slim matrix.
3. Pick one stack-or combine two. Examples:
   • Cesium for the globe plus Unity for building interiors.
   • Unreal in the cloud for a VIP demo while Babylon.js handles the day-to-day dashboard.

7. Take-away

The matrix stays; the scores shift. Unity solved our first urban twin, but in 2025, web-based options such as Cesium, and other cloud options close many gaps while Omniverse owns high-end engineering. Choosing the stack still starts with the clear criteria based on the project requirements.

And the call for standardization for data visualization across platforms remains apparent.

Even with newer engines, most teams still juggle different data-viz APIs, scene graphs, and file formats per platform. Industry groups are actively pushing standards (e.g., OGC 3D Tiles for geospatial streaming, Pixar USD for scene graphs, Khronos glTF for model interchange, OGTC efforts for Digital Twins), which underscores that the need for cross-platform data-visualization standards is both recognized and ongoing.