Spatial data describes things by where they are: a road, a border, a sensor reading, a satellite image. Geographic information systems store, analyse and map it. The central difficulty is that the Earth is a curved surface, so a position means nothing until you say which coordinate system it is in.
Key concepts
- Vector data: points, lines and polygons with attributes, such as a city, a river or a parcel.
- Raster data: a grid of cells holding values, such as elevation, temperature or an aerial image.
- Coordinate reference systems: latitude and longitude on a datum, and projections that flatten the globe with some distortion.
- Spatial operations: distance, intersection, buffer, containment and overlay.
- Spatial indexing: R-trees and grids that make a "what is nearby" query fast.
Vocabulary
- Projection: a mathematical method for drawing the curved Earth on a flat map.
- Geocoding: converting an address or place name into coordinates.
- GeoJSON, shapefile, GeoTIFF: common file formats for vector and raster data.
- Topology: how features connect and share boundaries, without regard to exact shape.
What to study, in order
- Coordinates, datums and projections: why the same place has different numbers in different systems.
- Open a vector dataset and a raster in a desktop GIS such as QGIS and style them.
- Use a spatial database with an extension such as PostGIS and run distance and containment queries.
- Write a script that reads GeoJSON and computes areas, distances and the nearest feature.
- Add a spatial index and measure the speed of a nearby-points query.
- Build a small web map from your own data, and check how projection choices change what it shows.
Related tags
#databases #computational-geometry #data-visualization #mobile-computing