Chapter 4: Geographical Information System (GIS)
In the last week of April 2019, a mass of cloud began to spin in the Bay of Bengal. The India Meteorological Department named it Fani and, day by day, drew its likely path — a curving line aimed at the coast of Odisha, in eastern India.
A line on a map, though, does not move people. What Odisha’s officials needed to know was more exact: which villages lay in the storm’s path, how many people lived in each, which roads could carry buses inland, and where the nearest cyclone shelters stood.
No single map holds all of that at once. So the pieces were kept as separate layers in a computer — the forecast track, the coastal population, the road network, the shelters — and laid one over another. Where the layers stacked, an answer appeared: these blocks of people, these routes, these buildings.
On that basis, before Fani struck near Puri on 3 May 2019, more than a million people were moved to safety. A cyclone that had hit the same coast in 1999 killed thousands; this time, planning built from stacked maps kept most people out of the storm’s reach.
The tool that lets you stack maps, put questions to them, and get a place-by-place reply is a Geographical Information System. This chapter is about how it holds the world as data — and what you can then ask of it.
Diagram
Fig. 4.0-1 Caption: The same place, held as separate layers that a GIS can lay one over another. Brief: Draw the classic GIS layer stack — four transparent rectangular map layers in receding perspective, one above the next, labelled from the top: storm track (a single curved line), coastal population (shaded blocks of varying density), road network (branching lines), cyclone shelters (a few points). Below the stack, one composite rectangle combining all four. Vertical arrows point downward from each layer into the composite, showing that the layers merge into a single integrated map. This is a concept diagram, not a map of any real area — do not draw a coastline or place names.
4.1 Spatial Information Technology
Almost every question in this book has been, at heart, a question about where. Where the monsoon breaks first, where people crowd together, where an industry finds its raw material. Spatial information technology is the family of tools built to answer such questions — to collect, store, retrieve, display, manage and analyse information that is tied to particular places on the earth. The word spatial means ‘relating to space’, that is, to location on the earth’s surface.
What makes this technology distinctive is that it never treats a fact on its own. It treats the fact together with the place it belongs to. A population figure is stored with the district it describes; a rainfall reading with the station that recorded it. Location is not a label added afterwards — it is built into the data.
Five Technologies, One Field
Spatial information technology is not a single invention but an amalgamation of five, each doing a different job. Four of them prepare and manage the data; the fifth brings them together for analysis.
Remote sensing is the gathering of information about the earth’s surface from a distance — from sensors carried on satellites or aircraft — without touching the ground being studied. The Global Positioning System (GPS) is a network of satellites that lets a receiver on the ground fix its own position, its latitude, longitude and height, to within a few metres.
Digital cartography is the making and storing of maps in digital form on a computer, in place of drawing them by hand on paper. A database management system (DBMS) is software that stores, orders and retrieves large quantities of data quickly and reliably. The fifth member, a geographical information system, is the tool that ties all of this together — storing data by its location, and letting you question and analyse it. It is the subject of the rest of this chapter.
Flowchart
Fig. 4.1-1 Caption: Five technologies feed spatial information handling; a GIS is where they are brought together. Nodes: Remote sensing (surface data from above) → GIS; GPS (precise ground positions) → GIS; Digital cartography (maps in digital form) → GIS; Database management system (orderly data storage) → GIS; GIS integrates the four → collect → store → manage → analyse → display spatial information
Seen this way, a GIS is not a rival to the other four but their meeting point. Remote sensing and GPS supply fresh data about the ground; digital cartography and a DBMS keep it in usable order; the GIS holds it all by location and answers questions from it. The next section pins down exactly what that fifth tool is, and why it can do what a folded paper map cannot.
4.2 What a GIS Is
The last section placed a GIS among its neighbouring technologies. This one pins down what it actually is, and then asks the sharper question: what can it do that a folded paper map cannot?
4.2.1 Defining a Geographical Information System
A geographical information system (GIS) is a computer system that stores data about places on the earth’s surface, together with the exact location of each place, and lets the user question, combine and analyse that data and display the results as maps. The core of the idea is that every piece of information is spatially referenced — tied to a known position on the earth by its coordinates.
That act of tying data to a real earth position is called georeferencing. Without it, a column of figures is only a list. With it, each figure has a place, and the system can ask how one place stands in relation to another.
There is no single agreed wording for GIS, and this is not sloppiness. The tools have changed so quickly that the definitions have changed with them, and each writer tends to define a GIS from their own field of work. What the working definitions all share is a list of core functions: a GIS captures data, stores it, checks it, integrates it, manipulates it, analyses it and displays it — always with the data referenced to the earth.
In practice, then, a GIS is two things joined together: a spatially referenced computer database, and the applications software that works upon it. The database holds the world as records; the software lets you interrogate them.
4.2.2 GIS Compared With Manual Map Methods
A paper map is a finished object. Which themes it shows, how they are coloured, what scale it is drawn at — all of this was decided before it was printed, and none of it can change without drawing a new map. A GIS, by contrast, keeps the data and builds the map on demand, so a single database can answer many different questions.
Three limits of the manual, paper method stand out. First, the information is processed and presented in one fixed way. Second, a map carries only the theme or themes chosen in advance — a soil map shows soils, not schools. Third, to change what is shown, or to ask something new, you must redraw the map from the beginning.
Against each limit a GIS offers a matching freedom. You can interrogate a feature on screen — select a district and read off the data attached to it — rather than only look at it. You can generate a fresh map by querying the data, for instance shading every district above a chosen density. And you can run spatial operations such as overlay or buffering, laying layers over one another (Fig. 4.0-1) to produce information that was present in none of the layers alone.
Most powerful of all is the last freedom. Because every record carries its location, data about one place gathered separately — its population held in one table, its rainfall in another — can be joined through that shared location. This linking of different data by place is the quiet engine of the whole system, and we return to how it is done later in the chapter.
4.3 The Two Forms of Geographical Information
To hold the world, a GIS must record two things about every feature: where it is, and what it is like. A river is somewhere on the earth, and it also has a width, a name and a rate of flow. These are the two forms of geographical information — spatial data and non-spatial data — and a GIS needs both to be useful.
4.3.1 Spatial Data
Spatial data answer the question ‘where?’. They record the position and shape of a feature on the earth’s surface. Whatever the feature — a town, a road, a lake — its spatial data fix it in place and give it a definite geometric form, so the system can store it and set it correctly against everything around it.
Point, Line and Area
Every spatial feature takes one of three forms. The difference is simply how much of the earth the feature occupies at the scale of the map.
A point feature has a single position and no real length or width at that scale. A well, a bus stop or a trigonometrical station is a point — the last being a surveyed ground marker whose exact position and height are known, one of the fixed reference points of a survey.
A line feature has length but negligible width, and is drawn as an ordered string of connected points. A road, a river, a railway or an electricity supply line is a line feature.
An area feature, also called a polygon, encloses space and is bounded by a closed line. A lake, a forest, a field or a district is an area feature.
Diagram
Fig. 4.3-1 Caption: The three forms of spatial data — a point, a line and an area. Brief: Three side-by-side panels. Left, a point feature: a small solid triangle labelled ‘trigonometrical station’. Middle, a line feature: a single curving line labelled ‘road’. Right, an area feature: a closed irregular outline with light shading or dots inside, labelled ‘forest’. A concept diagram, not a map of any real place — no coastline, no place names beyond these labels.
4.3.2 Non-spatial (Attribute) Data
Non-spatial data, also called attribute data, answer the question ‘what is it like?’. They describe a feature — its name, its size, its type, its value — but carry no location of their own. A point marks where a town is; the town’s name, population and literacy rate are its attributes.
Neither form is much use alone. A shape with no description is a blank outline: you know something is there but not what. A table of figures with no location cannot be placed on the ground and so cannot be mapped. The strength of a GIS is that it joins the two — every feature’s attributes are stored against its location, so the system knows at once both where a thing is and what it is like.
Take a river. Captured as a line, it has a spatial form. Stored against that line are its attributes: its name, its source and mouth, its width and its rate of flow. Ask the system about the river and it can return the shape, the description, or both — because it holds them together.
4.4 The Components of a GIS
It is tempting to think of a GIS as a piece of software. It is more than that. For a GIS to work, four components must come together — hardware, software, data and people. Remove any one and the system stops: software with no data has nothing to show, data with no one to question it just sits there.
Diagram
Fig. 4.4-1 Caption: The four components of a GIS, working as one system. Brief: A simple system diagram. At the base, a broad slab labelled ‘Hardware’. Resting on it, a box labelled ‘Software’. An arrow from a box labelled ‘Data’ (to one side) points into Software. At the top, a figure labelled ‘People’ with a downward arrow ‘operate’ into Software, and an upward arrow ‘results’ returning from Software to People. Keep it a clean concept diagram with the four labels clearly separated.
4.4.1 Hardware
Hardware is the physical machinery. At its centre is the computer itself, which does the processing and holds the storage. Around it sit the input devices that bring data in — a digitiser, a scanner, a GPS receiver, a keyboard — and the output devices that show or print the results — a monitor, a printer, or a plotter, which is a machine that draws large maps by moving a pen or print head across a wide sheet.
The scale of the hardware varies enormously. A GIS may run on a single laptop for a small study, or on multi-user supercomputers for a national mapping programme. The software is the same in kind; only the size of the machine changes.
4.4.2 Software
Software is the GIS program itself. Whatever the make, it must let the user carry out the whole chain of work: enter and edit data, store and manage it, transform it from one format or projection to another, query and analyse it, and finally display and output the results.
The managing of the stored data is done by a database management system built into or working alongside the GIS. Good software hides this machinery from the user, who sees only maps and answers, while the ordering of records goes on underneath.
4.4.3 Data
Of the four components, data are usually the most important — and the most expensive. A GIS handles spatial (geographical) data, and each item carries three things at once: its position, its connections with the features around it, and its non-spatial attributes.
Getting the spatial referencing right must be settled at the very start of a project, because everything built afterwards rests on it. Data may be bought from suppliers, captured by survey, or derived from remote sensing. In practice, gathering and cleaning the data often costs more than the hardware and software together.
4.4.4 People
No GIS runs itself. People plan the project, build and maintain the database, operate the software, and — most important of all — make decisions from what it produces. A GIS is only as good as the questions people put to it and the judgement they bring to its answers.
The number of people involved varies with the project. A small research study may have one person doing everything. A large organisation may split the work across teams — one capturing data, another analysing it, a third acting on the results.
4.5 Spatial Data Formats
We have seen that a GIS holds features as points, lines and areas with attributes attached. But at the lowest level, where the data are actually written into the computer, there are two rival ways of doing it: the raster format and the vector format. Almost every practical question about storing spatial data comes back to a choice between these two.
4.5.1 The Raster Data Format
Raster data store space as a grid — a mesh of rows and columns that divides the area into equal square cells, like the squares of graph paper. Each square is a grid cell, also called a pixel. Every cell is given a position, its row and column, and a single value standing for the attribute present there — a land-use code, an elevation, a colour.
The idea is easiest to see by hand. Draw a diagonal line on graph paper. To store it as raster, shade every square the line crosses. The line is now a pattern of filled and empty cells, and from the row-and-column position of each filled cell it can be rebuilt.
How faithful that rebuilt picture is depends on the cell size. Resolution is the relationship between cell size and the number of cells: smaller cells mean more cells for the same ground, and a finer, more detailed picture — but also far more data to store.
Diagram
Fig. 4.5-1 Caption: The same shape stored as raster at two resolutions — coarse cells and fine cells. Brief: Two grids side by side, each laid over the same simple blob-shape. Left: a coarse grid, roughly 10 columns by 8 rows, with the shape shown by fully shading each cell it covers, so the edge looks blocky; label ‘coarser cells — fewer, blockier’. Right: a finer grid, roughly 20 by 16, of the same shape, with a smoother shaded edge; label ‘finer cells — more, smoother’. Mark one square as ‘grid cell (pixel)’, and label the axes ‘rows’ and ‘columns’. A concept diagram, not a real map.
The cost of fine detail rises steeply. Take a square plot 100 m on each side. With cells 10 m across, it needs \(10 \times 10 = 100\) cells. Halve the cell size to 5 m and it needs \(20 \times 20 = 400\) cells. Halving the cell size has quadrupled the count — and the memory — while the plot itself is unchanged. This is why raster data are hungry for storage, and why compression is often applied.
Where Raster Fits
Raster suits some jobs and not others. It is the natural format for satellite images, scanned paper maps and aerial photographs, which arrive as grids of pixels already. It is chosen when very detailed images are wanted, when the cost of production must be kept low, when a plain ‘backdrop’ map is enough, or when individual features do not need to be picked out and analysed.
Its strengths follow from the grid. Raster is simple to understand and to program. Because every cell carries a value, it represents continuous, spread-out cover — land use, soil type, elevation — very well, and handles high spatial variability efficiently. Much source data from remote sensing come in raster form, so no conversion is needed, and most printers and plotters are raster devices.
Its weaknesses also follow from the grid. Coarse cells drop small features entirely. A cell that straddles two land uses must be generalised to one, blurring the true boundary. Line networks such as roads and rivers come out ragged. And each cell can hold only a single attribute, so several themes mean several separate grids.
4.5.2 The Vector Data Format
Vector data store space in a quite different way — not as filled cells, but as exact points pinned by their coordinates. A single point is written as two numbers, X and Y: X the distance from the left edge, Y the distance from the bottom edge. Where height matters, a third number, Z, is added for elevation above or depth below, making the data three-dimensional rather than two.
From points, everything else is built. A line is an ordered sequence of points joined one to the next, and the order gives the line a direction. An area, or polygon, is a sequence of points that closes back on itself. So a point, a line and an area are, underneath, all just lists of coordinates.
Diagram
Fig. 4.5-2 Caption: In the vector format, points, lines and areas are all stored as coordinate pairs. Brief: Three items on one panel. A single isolated dot tagged ‘X,Y’ and labelled ‘point’. A curving line drawn as a row of dots connected in order, several dots tagged ‘X,Y’, labelled ‘line’. A closed loop of dots connected into a polygon, tagged ‘X,Y’, labelled ‘area’. To one side, two reference arrows showing X measured from the left edge and Y from the bottom edge. A concept diagram, not a real map.
Because vectors hold exact coordinates, they can also record topology — the spatial relationships between features: which line joins which, which areas share a boundary, what encloses what. Topology is what lets a GIS answer ‘are these two roads connected?’ or ‘which plots border this one?’. Vector data are entered mainly by digitising, tracing features so the computer records the coordinates of each point, which the next section describes. Ground instruments such as total stations and GPS receivers also deliver positions directly as vector coordinates.
Vector suits precise work. It is chosen where high accuracy is needed, where individual features must be analysed, where file sizes must stay small, and where descriptive information is to be stored. It represents man-made features with clean edges — roads, plots, administrative boundaries — far better than a grid can, and its output is sharp rather than blocky. It is also efficient for topology-based analysis such as proximity and network analysis.
Its costs are the mirror of raster’s. Building topology is heavy work, and it must be rebuilt whenever the data change. The analysis algorithms are complex and demand powerful hardware for large datasets. Continuous surfaces such as elevation sit awkwardly in vector, coming out well only after values between points are interpolated. Overlay and filtering are harder to run within vector than within raster, and the software generally costs more.
Choosing Between Raster and Vector
Neither format is simply better; each fits a different kind of data. Raster is a grid of cells, at home with continuous, spread-out phenomena — satellite imagery, elevation, land cover — and simple to handle, but blocky at the edges and heavy on memory. Vector is a set of coordinates, at home with sharp-edged, discrete features — boundaries, plots, road networks — and precise and light for such work, but harder to build and clumsy for continuous surfaces.
The choice follows the job. Imagery and continuous surfaces call for raster; discrete features, boundaries and networks that need exact analysis call for vector. In practice many projects use both, converting from one to the other as the task demands.
4.6 The Sequence of GIS Activities
Building and using a GIS follows an order. Five activities run in sequence: getting the spatial data in, adding the attribute data, checking and correcting both, linking them together, and then analysing. The first four prepare the ground. The fifth — spatial analysis — is the payoff, and has a section of its own.
Flowchart
Fig. 4.6-1 Caption: The order in which a GIS is built and then put to work. Nodes: Spatial data input → Enter attribute data → Data verification and editing → Link spatial and attribute data → Spatial analysis (Section 4.7)
4.6.1 Spatial Data Input
Spatial data can enter a GIS in two broad ways: by acquiring ready-made digital data from a supplier, or by creating digital data yourself from maps and photographs.
Buying Ready-Made Data
Digital spatial data of every kind can now be bought from data suppliers, ranging from small-scale maps to large-scale plans. Public and private agencies are expected not to re-collect data that already exist. But when data come from more than one source, they must be checked for compatibility before being combined.
Compatibility turns on several things: the scale of the data, the georeferencing system used, the collection and sampling methods, the quality, the classification and interpolation methods, the size and shape of the mapping units, and the length of the record. Trouble arises most when datasets cross administrative boundaries — combining figures from two states or two districts, gathered by different methods, needs real care. Once the data are found compatible, they are transferred into the GIS on storage media or over a network.
Creating Data by Hand
Making the data yourself has four stages: enter the spatial data, enter the attribute data, verify and edit both, and link the spatial to the attribute. How it is done depends on whether the database is to be vector or raster. Three input devices do most of the work.
A digitiser is a flat tablet on which a map is fixed; the operator traces each feature with a hand-held cursor, and the tablet records the coordinates of every point traced — so it produces vector data. A scanner passes a light sensor across the whole sheet and turns it into a grid of pixels — so it produces raster data, a high-resolution image that may be used as it is or later processed into vector form. Scanners come in two kinds: those that record step by step, and those that capture the whole document in a single pass. A stereo plotter draws three-dimensional coordinates from overlapping aerial photographs, allowing height to be measured as well as position.
4.6.2 Entering the Attribute Data
Attribute data give the properties of each spatial feature. A river held as a line carries attributes stored against it — its width, its source and mouth, its rate of flow. These descriptions come from published records, official censuses, primary surveys or spreadsheets. They are entered by hand, or imported in bulk using a standard transfer format, and tied to the feature they describe.
4.6.3 Data Verification and Editing
Before the data can be trusted, they must be checked and corrected, usually with the computer’s help, so that errors are removed and accuracy is assured. Three errors commonly slip in during capture: spatial data that are incomplete or accidentally doubled, spatial data recorded at the wrong scale, and spatial data that are distorted.
This stage matters more than it looks. An error here is not local. Because every later map and analysis is built from this data, a single wrong scale or a bent boundary spreads quietly into every result drawn from it, where it is far harder to spot.
4.6.4 Spatial and Attribute Data Linkages
Linking must be done with care, because it is what makes a GIS powerful. A GIS can join different datasets that share a location. Suppose one table lists the population of each village, and another lists how far each village lies from the nearest primary health centre. Joined by village, the two together reveal which populous villages are also far from care — a fact present in neither table on its own.
There are three ways to match datasets, depending on how neatly they line up.
Exact matching is used when one file holds information about many features — a set of towns, say — and another file holds different information about the same towns. A key common to both, such as the town’s name, links them: each record carrying the same name is drawn from both files and joined into one.
Hierarchical matching handles data collected at different levels of detail. Some information is gathered often, over large areas; other information rarely, over small areas. When the small areas nest inside the larger ones, the small-area figures are added up until the grouped totals match the larger units, and only then are the two matched.
Fuzzy matching is needed when the boundaries themselves do not coincide. Crop boundaries, for instance, rarely follow soil boundaries. To find which soil grows a crop best, the two maps are laid one over the other and productivity worked out for each soil type. That laying of one map over another is exactly the operation the next section explains.
With the data now in, checked and linked, the GIS is finally ready for its real purpose — to be questioned.
4.7 Spatial Analysis
Storing and drawing maps is useful, but it is not what makes a GIS special — a filing cabinet stores and a drawing program draws. What sets a GIS apart is spatial analysis: the ability to put questions to the data about how places relate, and get back new information that no single map held.
A GIS offers several analytical operations. The four main ones are overlay, buffer analysis, network analysis and the Digital Terrain Model. Network analysis works on connected line features — roads, pipes, wires — to find shortest routes and test connectivity. A Digital Terrain Model represents the height of the ground surface, from which slope, drainage and visibility can be worked out. This section treats the two most important operations in full — overlay and buffer — and leaves the other two named.
4.7.1 Overlay Operations
Overlay lays two or more thematic layers of the same area one on top of another to make a new layer that combines them. It works like sieve mapping, the old practice of laying tracing-paper maps on a light table to see where features coincide — only now the computer does the lining up, exactly and instantly.
The simplest overlay is a union, written as x + y = z. A road map laid over a rail map yields a single transport map that carries both. The output holds something neither input did on its own: where roads and rails cross, where a soil type and a crop coincide, where a flood zone meets a settlement.
Diagram
Fig. 4.7-1 Caption: An overlay combines two layers of the same area into a new one. Brief: Three panels left to right, joined by a ‘+’ and an ‘=’. Panel 1, labelled ‘Road map’: a few curving roads. Then ‘+’. Panel 2, labelled ‘Rail map’: a single rail line drawn with the standard cross-tick railway symbol. Then ‘=’. Panel 3, labelled ‘Transport map’: the roads and the rail line together in one frame. Below, the operation ‘overlay (union): x + y = z’. A concept diagram, not a map of any real area.
Because the output combines layers, overlay is the tool for studying change. Lay an old land-use layer over a recent one and the differences show what changed and where. The same move, applied to satellite images from different dates, is used to measure land transformation, the retreat of glaciers, and the ground-level effects of a warming climate.
4.7.2 Buffer Operations
Buffer analysis draws a zone of a chosen distance around a feature — a point, a line or an area — to find what falls inside it. A 500 m buffer around a factory, a 500 m buffer along a road, a 1 km buffer around a school: each marks out the ground within that reach, so the GIS can pick out whatever lies inside.
Diagram
Fig. 4.7-2 Caption: A buffer is a zone of fixed distance drawn around a point, a line or an area. Brief: Three stacked rows, each showing a feature on the left and its buffered version on the right, joined by an arrow. Row 1: a single dot, arrow to the same dot surrounded by a shaded circle, labelled ‘buffer around a point (e.g. 500 m)’. Row 2: a curving line, arrow to the same line flanked by a shaded band on both sides, labelled ‘buffer around a line (e.g. 500 m)’. Row 3: a small filled rectangle, arrow to the same rectangle surrounded by a shaded margin, labelled ‘buffer around an area (e.g. 1 km)’. A concept diagram, not a real map.
Buffers answer questions of access and of influence. Around a facility — a hospital, a school, a post office — a buffer shows who is served and who is left out. Around a source of harm — of air, water, land or noise pollution — it shows who is exposed. Planners routinely draw buffers of 2, 4, 6, 8 or 10 km around towns with major health or education centres to see how far the benefit reaches; as a rule, places inside are better served than those outside.
A buffer count is only as honest as its assumption — that straight-line distance equals access. A village 4 km away as the crow flies may lie across a river with no bridge, or up a hillside with no road, and so sit inside the circle yet far from real help. The GIS measures distance on the map, not the journey on the ground. Used well, buffer analysis shows planners where facilities are genuinely missing; used carelessly, it can declare people served who are not.
Overlay and buffer are only two of the operations a GIS offers, but together they show the essential move of spatial analysis: turn a question about places into a stack or a shape, and read the answer off the combination.
4.8 Summary
- Spatial information technology is the family of tools for handling information tied to place. A GIS is the one that brings the others — remote sensing, GPS, digital cartography and a database system — together for analysis.
- A GIS is a computer system that stores georeferenced data and lets you question, combine, analyse and display it. In practice it is a spatially referenced database joined to the software that works upon it.
- Its advantage over a paper map is that the map is never fixed. The GIS stores the data and draws the map on demand, so one database can answer many questions, combine layers, and link records through their shared location.
- Geographical information takes two forms. Spatial data say where a feature is and what shape it holds — a point, a line or an area. Attribute data say what it is like. A GIS keeps the two together.
- Four components make a GIS work: hardware, software, data and people. Data are usually the costliest, and must be correctly referenced from the outset; people frame the questions and judge the answers.
- Spatial data are written in two formats. Raster stores space as a grid of cells — good for imagery and continuous surfaces, but blocky and heavy on memory. Vector stores it as coordinate points — precise and light for sharp-edged features, but awkward for continuous surfaces. The job decides the format.
- Building a GIS runs in sequence: input the spatial data (a digitiser yields vector, a scanner yields raster), enter the attributes, verify and edit — because an error here spreads into every later result — then link spatial to attribute data by exact, hierarchical or fuzzy matching.
- Spatial analysis is the point of it all. Overlay stacks layers of one area into a new combined layer, showing where features coincide and how they have changed. Buffer draws a zone of set distance around a feature to show who is served, or exposed. Network analysis and the Digital Terrain Model are two further operations.
- Every such answer carries a limit. A buffer measures distance on the map, not the real journey on the ground; an overlay trusts that its layers truly line up.
Flowchart
Fig. 4.8-1 Caption: How a GIS turns the world into data and reads an answer back out of it. Nodes: Real-world features → captured as spatial data: point; line; area → plus attribute data (the description) → stored in one format: raster (grid of cells) or vector (coordinate points) → verified and edited → linked by shared location → analysed: overlay (stack layers into a new one); buffer (zone of set distance around a feature) → a new map or a place-by-place answer → a decision
What This Chapter Cannot Tell You
A GIS is only as good as the data poured into it. A clean-looking map may rest on features captured at the wrong scale or already out of date, and the map will not betray the flaw.
Its answers inherit their assumptions. A buffer treats straight-line distance as access; an overlay treats two layers as perfectly aligned. Where an assumption fails, the answer misleads while still looking authoritative.
And a GIS shows only what has been put into it. A feature that was never captured is, as far as the system is concerned, simply not there.
4.9 Key Terms
- Area feature — a spatial feature that encloses space, bounded by a closed line, such as a lake or a district; also called a polygon.
- Attribute data — data describing a feature (its name, size, type or value) but carrying no location of their own.
- Buffer — a spatial-analysis operation that draws a zone of set distance around a point, line or area to find what falls inside it.
- Database management system (DBMS) — software that stores, orders and retrieves large quantities of data quickly and reliably.
- Digital cartography — the making and storing of maps in digital form on a computer rather than by hand on paper.
- Digitiser — a flat tablet on which a map is traced by hand, recording the coordinates of each point as vector data.
- Digitising — capturing spatial data by tracing features so that the computer records the coordinates of each point.
- Exact matching — linking two datasets about the same features through a key common to both, such as a shared name.
- Fuzzy matching — combining datasets whose boundaries do not coincide by overlaying them and reading off each combination.
- Geographical information system (GIS) — a computer system that stores georeferenced data and lets the user question, combine, analyse and display it.
- Georeferencing — the tying of data to a known position on the earth by its coordinates.
- Global Positioning System (GPS) — a network of satellites that lets a receiver fix its own position on the earth to within a few metres.
- Grid cell — a single square of a raster grid, holding one attribute value; also called a pixel.
- Hierarchical matching — combining datasets collected at different levels of detail by adding small-area figures up to match larger units.
- Line feature — a spatial feature with length but negligible width, stored as an ordered string of points, such as a road or river.
- Non-spatial data — data describing a feature but carrying no location of their own; also called attribute data.
- Overlay — a spatial-analysis operation that lays two or more layers of the same area together to make a new combined layer.
- Pixel — a single grid cell of a raster, holding one attribute value.
- Point feature — a spatial feature at a single position, with no real length or width at the map’s scale, such as a well or bus stop.
- Raster data — spatial data stored as a grid of equal cells, each cell holding a single value.
- Remote sensing — gathering information about the earth’s surface from a distance, by sensors on satellites or aircraft, without touching it.
- Resolution — the relationship between cell size and the number of cells in a raster; smaller cells give finer detail.
- Scanner — a device that converts a whole map into a grid of pixels, producing raster data.
- Spatial data — data recording where a feature is and what shape it has: a point, a line or an area.
- Spatial information technology — the family of technologies for collecting, storing, managing and analysing information tied to location.
- Stereo plotter — a device that draws three-dimensional coordinates from overlapping aerial photographs.
- Topology — the spatial relationships between features: which connects to which, which share a boundary, and what encloses what.
- Vector data — spatial data stored as exact points defined by their coordinates, with lines and areas built from ordered points.
4.10 Points to Ponder
A printed map is finished the moment it is drawn; a GIS map is remade every time it is asked a question. Does that make the GIS map more truthful, or only more changeable?
The same city is a point on a map of India and an area on a map of its own district. If a feature’s very shape depends on the scale you choose, is the shape a property of the place, or of the map?
An overlay can reveal where a flood zone meets a settlement — a fact that was present in neither map alone. Where does that new fact come from, if it was in none of the inputs?
A buffer counts a village as ‘served’ because it lies within 5 km of a hospital, even if a bridgeless river stands between them. When does a number that is technically correct become practically false — and can a GIS ever notice the difference on its own?
When a buffer decides who is inside a service zone and who is outside, a line on a screen becomes a line in people’s lives. Who should choose that distance — the technician who can draw it, the planner who reads it, or the people it divides?
4.11 Objective Questions
- Which of the following would be stored as attribute data rather than spatial data?
- the shape of a district
- the location of a well
- the literacy rate of a state
- the course of a river [1 mark]
- In the vector data format, a line is stored as —
- a single grid cell
- an ordered sequence of coordinate points
- a shaded block of pixels
- a single X, Y pair [1 mark]
- The relationship between cell size and the number of cells in a raster is called its —
- topology
- resolution
- georeferencing
- attribute [1 mark]
- Which spatial-analysis operation would you use to find every settlement within 3 km of a river?
- overlay
- network analysis
- buffer
- digitising [1 mark]
- Assertion (A): A GIS can answer questions that a printed paper map of the same area cannot. Reason (R): A GIS stores the data and generates maps on demand, rather than fixing them in print.
- Both A and R are true and R is the correct explanation of A.
- Both A and R are true but R is not the correct explanation of A.
- A is true but R is false.
- A is false but R is true. [1 mark]
- Assertion (A): Raster data represent line networks such as roads more accurately than vector data. Reason (R): Raster stores space as a grid of equal cells.
- Both A and R are true and R is the correct explanation of A.
- Both A and R are true but R is not the correct explanation of A.
- A is true but R is false.
- A is false but R is true. [1 mark]
- Match each item in Column A with the correct entry in Column B.
| Column A | Column B |
|---|---|
| a. Digitiser | i. raster data |
| b. Scanner | ii. spatial relationships between features |
| c. Stereo plotter | iii. vector data |
| d. Topology | iv. three-dimensional coordinates from aerial photographs |
[1 mark]
Fill in the blank: The four components of a GIS are hardware, software, data and __________. [1 mark]
Fill in the blank: Tying a piece of data to a known position on the earth by its coordinates is called __________. [1 mark]
State whether true or false: In the raster format, each grid cell can store more than one attribute value at a time. [1 mark]
4.12 Check Your Concepts
Define a geographical information system (GIS). [2 marks]
State the three geometric forms that spatial data can take, giving one example of each. [2 marks]
Differentiate between spatial data and attribute data. [3 marks]
Give reasons why data are usually the most important and the most expensive component of a GIS. [3 marks]
Explain how a buffer operation works, using the example of siting a facility such as a school or a hospital. [4 marks]
‘A GIS map is always more reliable than a paper map.’ Evaluate this claim. [4 marks]
Differentiate between the raster and vector data formats, and state one task for which each is better suited. [4 marks]
Describe the sequence of activities involved in building and using a GIS, from spatial data input through to spatial analysis, explaining why verification and linkage matter. [6 marks]
Overlay and buffer are the two most important operations of spatial analysis. Describe each with an example, and explain what makes spatial analysis the capability that distinguishes a GIS from an ordinary database. [6 marks]
4.13 Map, Diagram & Data Interpretation
- Study Fig. 4.7-1, the overlay diagram.
- Name the operation shown and write it in the form x + y = z. [1]
- State one piece of information present in the output map that was in neither input map. [1]
- Give one real use of overlay for studying change over time. [1] [3 marks]
- Look at Fig. 4.5-1, which shows the same shape stored as raster at two resolutions.
- Which grid stores more detail, and why? [1]
- State what happens to a feature that is smaller than one cell. [1]
- A square plot 200 m on each side is divided into cells 20 m across. How many cells are needed? Show your working. [2] [4 marks]
- A GIS is asked to find the population within reach of a fire station, using a service radius of 4 km.
- Which spatial operation is this? [1]
- Calculate the area of the circular service zone, using area = π r². Show each step and give your answer in km². [2]
- State one reason a village lying inside this zone might still not be truly served. [1] [4 marks]
- Study Fig. 4.6-1, the sequence of GIS activities.
- List the five activities in the order shown. [1]
- Explain why an error made at the verification and editing stage is more serious than it first appears. [2] [3 marks]