Base Maps: Survey of India & OpenStreetMap

A base map is the reference layer you draw everything else on top of: the roads, rivers, coastline, built-up areas and terrain that let a viewer (and you) locate the plot in the real world. It is not usually the thing you analyse. It is the thing that makes your analysis legible.

The base map you pick quietly decides three things: how accurate your positions are, what you are allowed to do with the result, and how current the map is. An official surveyed sheet is positionally authoritative but may be a decade old and cannot be used commercially without permission. A crowd-sourced map is astonishingly current in a busy city and free to build on, but its accuracy varies street by street. Knowing which trade-off you are making is the whole skill.

The Survey of India (SoI) is the country's national surveying and mapping agency, under the Department of Science and Technology. Its Open Series Maps (OSM) — confusingly sharing the initials of OpenStreetMap, so we will say Open Series Maps in full — are topographic sheets at 1:50,000 scale covering the whole country, roughly 3,000 sheets. You download them as PDFs from the Nakshe web portal.

Two honest caveats. First, access is Aadhaar-gated: the portal authenticates Indian residents before it releases a free sheet, and the licence is non-commercial. Second, a PDF topo sheet is a picture of a map, not GIS-ready vector data. To use it in analysis you must georeference and, if you want features, digitise it — the exact craft of Module 3. What SoI gives you is authority: contours, spot heights, settlement names and drainage that were actually surveyed, which is the reference you check crowd data against.

OpenStreetMap (OSM) is a global, crowd-sourced vector map — think of it as the Wikipedia of geography. Volunteers and organisations trace roads, buildings, water, land use and points of interest, and anyone can download the result as real vector features with attributes. In dense Indian cities the coverage of roads and amenities is often excellent and remarkably up to date; in a small village it can be sparse.

OSM is licensed under the Open Database License (ODbL 1.0): free to use, including commercially, provided you attribute OpenStreetMap and share any improvements to the data under the same terms. That share-alike clause matters for planners publishing derived datasets, so read it before you build a product on it. The great gift of OSM is that it is already vector: no georeferencing, no digitising — you query it and it arrives as clickable, queryable features.

OSM does not have neat columns like a survey database. Instead every feature carries free-form tags in key=value pairs: a road is a line tagged highway=primary; a hospital is a point or polygon tagged amenity=hospital; a building is a polygon tagged building=yes. When you download OSM into QGIS, those tags become the attribute table you query.

This is liberating and dangerous in equal measure. Liberating, because you can pull exactly the theme you need — every school, every bus stop, every drain — with one query. Dangerous, because tagging is done by volunteers and is uneven: some buildings carry a height, most do not; a road tagged residential in one city may be tagged unclassified in another. Treat OSM tags as a rich but incomplete index, never as a legally authoritative record.

In practice most projects run both. You start on OSM because it is instant vector context — roads, blocks, amenities — that lets you begin analysis the same afternoon. You keep the Survey of India sheet (and ISRO Bhuvan imagery and thematic layers) as the authority you validate against: does the OSM road network match the surveyed alignment, is the drainage really where the crowd drew it, is a contour plausible.

For a statutory deliverable, the hierarchy is clear: official surveyed data and the municipal or development-authority base outrank crowd data. For a fast feasibility study or a studio concept, OSM is often enough. The professional move is to state which base you used and why on the map itself, so anyone reading it knows how far to trust each line.