Web GIS: Leaflet, Mapbox & Earth Engine

A desktop GIS redraws the whole map from your data every time. A web map cannot afford that, so it uses a tile trick. The map is pre-cut into small square images or data chunks at many zoom levels - a tile pyramid: one tile covers the world at zoom 0, four at zoom 1, sixteen at zoom 2, and so on down to street level. When you pan or zoom, the browser fetches only the handful of tiles currently on screen. This is why a slippy web map feels instant even over a phone connection - it never loads more than it needs.

The address of a tile is just its zoom, column and row (the {z}/{x}/{y} you see in an XYZ URL). You already met these in QGIS as XYZ Tiles. Understanding the pyramid is the key to everything else in web GIS: performance, cost and the raster-versus-vector-tile choice all flow from it.

Tiles come in two kinds, and the difference shapes what your map can do.

Raster tiles are pre-rendered pictures - the styling is baked into the pixels on a server. They are simple, work everywhere, and are what lightweight libraries usually serve. But you cannot restyle them in the browser (dark mode means a whole new tile set) and you cannot click a feature to read its data, because there are no features - just an image.

Vector tiles send the actual geometry - points, lines, polygons - and let the browser draw and style them live, using the graphics card (WebGL). Now the same data can be recoloured instantly, labels can rotate as you do, and features are clickable and queryable. The trade-off is more work for the browser and a more complex setup. As a rule: raster tiles for a simple, robust backdrop; vector tiles when the map must be interactive, restylable or data-rich.

A web map is drawn by a JavaScript library. Four matter.

Leaflet is the small, friendly one - lightweight, easy, and the usual first choice for a simple interactive map with markers and raster tiles. Mapbox GL JS renders vector tiles beautifully on the GPU, but it is a commercial product: v2 and later use a proprietary licence and require a Mapbox account and access token, with a free tier and paid usage beyond it. MapLibre GL JS is the open-source fork of Mapbox GL - the same vector-tile power, free and unencumbered, and increasingly the default for teams who want GPU vector maps without vendor lock-in. OpenLayers is the heavyweight: feature-rich, handles OGC services (WMS, WFS, WMTS) and reprojection out of the box, at the cost of a steeper learning curve.

For large data-visualisation overlays there are deck.gl and the no-code kepler.gl built on it. You do not need to master any of these to use web GIS - ArcGIS Online and QGIS-to-web plugins generate web maps for you - but knowing the names lets you brief a developer and read what a tool is doing.

Web GIS is not only about display. Google Earth Engine (GEE) is a cloud platform that pairs a petabyte-scale catalog of satellite and geospatial data with the compute to analyse it - free for research, education and non-commercial use (commercial use is paid). The idea inverts the normal workflow: instead of downloading gigabytes of imagery to your machine, you send a short script up to where the data already lives, and only the answer comes back.

That makes previously impossible questions routine. Map vegetation change across an entire metropolitan region over ten years; detect built-up sprawl for every city in a state; build a cloud-free composite from years of Sentinel and Landsat passes - all in minutes, in a browser, on hardware you do not own. For an Indian planner or researcher without a supercomputer, GEE is quietly transformative. We use it properly in the remote-sensing module; for now, register the shift: the cloud can be the computer, not just the delivery van.

Reach for web GIS when the audience will not install software (clients, committees, the public); when the map must be live, interactive and shareable by a link; when many people need one always-current map; or when the analysis is too big for a laptop and belongs in the cloud (Earth Engine).

Stay on the desktop when you need heavy, precise geoprocessing and full control; when you are editing data carefully; when the work is confidential and must not leave your machine; or simply when the internet is unreliable - a real constraint on many Indian sites. The mature workflow uses both: analyse and prepare on the desktop (QGIS or ArcGIS Pro), then publish to the web for reach. Web GIS is a distribution and cloud-compute layer on top of your GIS skills, not a replacement for them.