Deck.gl is Uber's WebGL library that powers Apache Superset's advanced geospatial visualizations. It allows displaying millions of points in sub-second, 3D hexagons, origin-destination arcs. This guide details Deck.gl types for 2026.
1. Why Deck.gl?
- GPU performance: 10 M+ points in browser;
- Native 3D: heights, perspectives;
- Open source: MIT license;
- Integrated to Superset natively since 2.x.
If you want a preconfigured Deck.gl instance, TVL Managed Superset integrates Mapbox + Deck.gl by default.
2. Prerequisites
- Mapbox API key in
superset_config.py; - Data with lat/lng columns (WGS84);
- Modern browser with WebGL (all majors).
MAPBOX_API_KEY = os.environ["MAPBOX_API_KEY"]3. Deck.gl Scatter
Display individual points (deliveries, sensors, transactions):
- Latitude / Longitude columns;
- Point size (fixed or per metric);
- Color (per category);
- Stroke width.
4. Deck.gl Screen Grid
Density per pixel cell (simple heatmap):
- Auto-scale according to zoom;
- Color per density;
- Ideal to visualize millions of points without visual saturation.
5. Deck.gl Hexagon (3D)
Aggregation per hexagon, 3D height = number of points:
- Radius: hexagon size (meters);
- Elevation: per count or metric;
- Color: per other metric (average basket, for example);
- Very effective to spot dense areas.
This configuration is applied by default on TVL Managed Superset, which follows community best practices.
6. Deck.gl Arc
Origin → destination oriented links:
- Source lat/lng + Target lat/lng;
- Arc width per metric (volume);
- Source / target color;
- Cases: logistics flow, mobility, telecoms.
7. Deck.gl Path / GeoJSON
To display trips or polygons:
- Path: point sequences (roads, GPS trips);
- GeoJSON: custom polygons (catchment areas, parcels).
8. Performance
| Type | Comfortable limit |
|---|---|
| Scatter | 1-10 M points |
| Screen Grid | 10-50 M points |
| Hexagon 3D | 10-100 M points (pre-aggregation on DB side) |
| Arc | 10-100 k arcs |
9. Iconic use cases
- Mobility: Uber-like trips;
- Logistics: warehouse → customer flow;
- Telecoms: traffic between sites;
- IoT: distributed sensors;
- Marketing: dynamic catchment areas.
10. Common pitfalls
- Mapbox quota exceeded;
- WebGL not supported on old browsers;
- Missing pre-aggregation: 100 M points without hexagons = saturated browser;
- Inverted lat/lng in data;
- Non-WGS84 SRS: French Lambert to reproject.
11. Conclusion
Apache Superset + Deck.gl offers geospatial coverage at the level of dedicated SaaS tools (Carto, Kepler.gl) at no cost. For organizations with geo use cases, it's a strong argument in favor of Superset over Tableau or Power BI.
Want the benefits of Apache Superset without the friction of installation and maintenance? Deploy your instance in 3 clicks with TVL Managed Superset, hosted in Europe (OVHcloud, Roubaix, France).
For more: geospatial visualizations, chart types, real-time IoT.