Geographic Information Systems and Spatial Data Fundamentals
This domain tests whether you understand what makes spatial data spatial: geometry plus location plus attributes, and the consequences of that combination for storage, analysis and display. Items blend definitions with short reasoning scenarios, so you must be able to pick the right model and spot a broken assumption.
The components of a GIS
A working GIS is described as hardware, software, data, people and methods/procedures. Exam items often describe a project failure and ask which component was neglected; the answer is rarely hardware, it is usually data (wrong datum, no metadata) or methods (undocumented workflow).
Vector and raster
- Vector models represent discrete entities as points, lines and polygons; boundaries are crisp, attributes are stored per feature in an attribute table, and topology can be built.
- Raster models represent continuous variation as a grid in which each cell carries one value. Cell size defines spatial resolution: a 10 m raster holds 100 cells per hectare.
- The phenomenon drives the choice, not preference: land cover, elevation and suitability surfaces are naturally raster; parcels, pipes and roads are naturally vector.
Three meanings of scale
Cartographic scale (the map ratio, 1:50,000), geographic extent (a neighbourhood versus a continent) and analysis resolution are distinct concepts. 1:10,000 is a larger scale than 1:250,000 because it shows a smaller area with more detail. "Large scale" never means "large area".
Worked example
A team must report the share of building floor area lying within 300 m of a transit stop. Buildings are polygons with 48,000 records and stops are points, both stored in EPSG:4326. Correct reasoning: reproject to a projected CRS in metres, buffer the stops at 300 m, intersect the buffer with the buildings, then divide the intersected polygon area by total building area. Buffering 300 "units" in EPSG:4326 would create a 300-degree buffer and a meaningless answer. A spatial join alone only tags buildings with stop attributes; it does not measure partial overlap, so an area-weighted answer needs an overlay or clip.
Traps that cost marks
- Believing a spatial join edits geometry; it only transfers attributes.
- Treating decimal degrees as metric distances.
- Confusing precision (number of digits) with accuracy (closeness to truth).
- Assuming finer resolution is always better: smaller cells cost storage and may exceed the accuracy of the source data.