GIS & Urban Data

A paper map is a picture; a Geographic Information System is a database that happens to draw pictures. That distinction is the whole lesson in one sentence. When you open a GIS project you are not looking at a single fixed image but at a stack of independent layers, each holding a different kind of geographic thing, roads, buildings, wards, water, trees, bus stops, and each of those things carries a table of attributes behind it. A road is not just a line; it is a line that also knows its name, its width, its surface, its traffic count. Because every feature is both a shape and a record, you can interrogate the map the way you interrogate a spreadsheet: show me every plot within three hundred metres of a metro station, or every ward where density exceeds twenty thousand people per square kilometre, or every stretch of road without a footpath. The software then draws the answer. This is the leap that matters for a designer. A survey drawing tells you what is there; a GIS lets you ask what the arrangement means. Studio Matrx runs a full, dedicated GIS course for those who want to master the software, so this lesson stays deliberately at the designer's altitude: enough to make you dangerous with data, not enough to make you a technician. Our goal is that you can commission, read, and sanity-check spatial analysis, and do a fair amount yourself, without pretending to be a remote-sensing specialist.

All spatial data comes in one of two flavours, and knowing which you are holding saves endless confusion. Vector data represents the world as discrete geometry: points, lines and polygons. A tree is a point, a road a line, a plot or a ward a polygon. Vectors are crisp at any zoom, small in file size, and carry rich attribute tables, so they are what you use for anything with a clear boundary or identity, cadastral parcels, administrative wards, the street network, land-use zones. The common vector formats you will meet are the old but universal shapefile and the cleaner, web-friendly GeoJSON. Raster data, by contrast, represents the world as a grid of cells, like a photograph, where each pixel holds a value. A satellite image is a raster; so is a digital elevation model where each cell stores a height, or a heat map where each cell stores a temperature. Rasters are how you handle continuous phenomena, terrain, temperature, rainfall, vegetation, light, that do not have neat edges. A useful rule of thumb: if the thing has a boundary you could walk along, store it as vector; if it varies smoothly across space, store it as raster. Real urban analysis constantly moves between the two. You might drape building footprints (vector) over a slope model (raster) to find plots too steep to build, or count how many green pixels (raster) fall inside each ward polygon (vector) to compare neighbourhood tree cover. The figure shows the same street corner in both models so the difference lands visually.

The power of GIS is that layers are independent yet perfectly aligned. You can switch the parks layer on, the traffic layer off, restyle the wards by population, and reorder what sits on top, all without touching the underlying data. What keeps every layer honest is the coordinate reference system, the CRS, the mathematical scaffolding that says where on the round earth a flat coordinate actually falls. The single most common beginner failure is mixing layers in different coordinate systems, so that your roads and your buildings, both correct, sit hundreds of metres apart or, worse, land in the ocean off West Africa where the zero-zero origin lives. For India you will usually work in a geographic system based on WGS 84 for global data or a projected system such as the relevant UTM zone when you need to measure real distances and areas in metres. Underneath your working layers sits the base map, the contextual backdrop, an aerial image, an OpenStreetMap tile layer, or a plain grey canvas, that orients the eye without adding data you must manage. A disciplined GIS habit is to name and group layers clearly, keep one project CRS, and treat the base map as scenery, not evidence. The figure renders this as a layer cake: the base map at the bottom, then terrain, then blocks and streets, then land use, then people, each transparent sheet contributing one dimension of the city to the composite you finally read.

A GIS is only as good as the data poured into it, and for the Indian city the richest single source is the Census of India, which reports population, households, literacy, work and amenities down to the ward and even enumeration-block level. Join that census table to a layer of ward boundaries and the city suddenly reveals its demographic geography, where the young live, where households are largest, where basic services thin out. Beyond the census, a growing constellation of open data exists: national and municipal open-data portals, the Smart Cities and AMRUT programmes that pushed many cities to publish spatial data, transit authorities releasing route and stop data, and utility maps of varying quality. The honest caveat is that Indian spatial data is uneven, cadastral records are often not digitised or not public, ward boundaries change between censuses, and a layer may be years stale, so a professional always records the source, the vintage and the licence of every layer, and never presents a stale map as current truth. Two disciplines separate the amateur from the professional here. First, provenance: know exactly where each layer came from and when. Second, joinability: data is only useful if it shares a key, a ward code, a plot number, a place name, with the geometry you want to attach it to. Much of real GIS work is the unglamorous labour of cleaning names and codes so that a spreadsheet from one office will marry the boundary file from another.

For a designer starting a project in an Indian city today, OpenStreetMap is often the fastest route to a workable base of streets, buildings, water and amenities, and it deserves a section of its own. OpenStreetMap is a free, editable map of the world built by a global community of volunteers, the Wikipedia of geography, and its data is genuinely open to download and use. In many Indian cities its coverage of the road network, points of interest and building footprints now rivals or exceeds commercial sources, precisely because local mappers know the lanes that no survey recorded. You can pull OpenStreetMap data straight into QGIS, extract just the layers you need, roads, footpaths, shops, parks, and have a credible base map for analysis in an afternoon. Two cautions temper the enthusiasm. First, quality is uneven and unverified: because anyone can edit, coverage is superb in some neighbourhoods and thin in others, and a designer must ground-truth what matters. Second, completeness varies by feature: main roads are excellent, but footpaths, informal settlements and small lanes may be under-mapped, exactly the fabric an urban designer cares about most. The constructive response is to contribute back, adding the missing footpath or corrected street name you discover on site, so the commons you drew from grows richer. OpenStreetMap embodies a quietly radical idea, that the map of a city need not be the property of a state or a company but can belong to the people who live in it.

Some of the most powerful urban data now falls from the sky. Remote sensing, the science of measuring the earth from satellites and aircraft, gives the urban designer a synoptic, repeatable and comparable view no ground survey can match. At its simplest, high-resolution satellite imagery is a base map you can trace, reading built form, open space, tree canopy, water bodies and the tell-tale texture of informal settlements straight off the picture. But remote sensing goes far beyond pretty pictures, because satellites see wavelengths the eye cannot. Near-infrared bands reveal healthy vegetation, letting analysts compute a greenness index and map the city's living cover objectively. Thermal bands reveal surface temperature, exposing the urban heat island as glowing red arteries of tarmac and roof against the cool of parks and lakes, a decisive input for climate-responsive design. Because the same satellite passes overhead repeatedly, you can also detect change: comparing images across a decade shows a lake shrinking, a wetland being built over, a city's edge sprawling into farmland, evidence that is hard to argue with in a public hearing. For the designer, the discipline is to treat imagery as evidence, not decoration: note the date and resolution, understand that a shadow is not a wall and a blue roof is not a swimming pool, and confirm the critical readings on the ground. Used well, the view from above turns a site visit into a diagnosis of the whole territory the site belongs to.

The good news for the not-for-profit-minded and the student alike is that the professional-grade tool is free. QGIS is a mature, open-source desktop GIS that does almost everything the expensive commercial packages do, and it has become the lingua franca of planning studios and NGOs across India. You do not need to master it to benefit from it; a handful of operations covers most of what design work demands. The spatial join marries two layers by location, tagging each bus stop with the ward it falls in, or counting how many schools sit inside each neighbourhood polygon. The buffer draws a zone of a given radius around features, the classic four-hundred-metre walk-shed around a station or a park, so you can see who it actually serves. The clip and intersect trim layers to a study boundary or find where two conditions overlap, say, flood-prone land that is also zoned residential. Thematic styling, colouring wards by density or plots by land use, turns a table into an argument the eye grasps instantly. And simple measurement, of area, length and count, answers the questions clients ask first: how much open space per person, how many metres of street without a footpath, how many homes within walking distance of transit. None of this requires programming; all of it requires clear questions. The professional habit is to start from the design question, choose the smallest analysis that answers it, and keep the map honest, one message, sourced layers, a legend a councillor could read, rather than a rainbow of data for its own sake.