Module 1: Introduction to GIS

1.3 The Five Components of a GIS

Hardware, software, data, methods, and people — the canonical five components and why each matters.

Lesson 3 of 100·18 min read

Key takeaways

A GIS is a socio-technical system: data, software, hardware, procedures, and people are all first-class components.

Weakness in any one component limits the value of the whole system.

Modern deployments shift the balance — hardware matters less, data and methods matter more.

Introduction

The classic GIS textbooks describe five components. Sometimes you see six (adding network), sometimes four (folding methods into software). The exact count matters less than the underlying idea: a GIS is not a single product — it's a system, and systems fail at their weakest link.

This lesson walks through each component, shows how modern shifts have reshaped them, and explains what to watch for when you're setting up or evaluating a GIS deployment.

1. Data

Data is the substrate of every GIS. Without data, the software is an empty calculator. There are two data families:

Vector data — points, lines, polygons. Examples: road centrelines, parcel boundaries, sample locations.
Raster data — regular grids of cells, each carrying a numeric value. Examples: satellite imagery, elevation models, temperature surfaces.

In addition, GIS uses attribute data (tables describing features), metadata (data about data), and network data (topological graphs for routing).

Data type	Typical example	Typical sizes
Vector points	GPS positions	kilobytes to gigabytes
Vector lines	Road networks	megabytes to tens of gigabytes
Vector polygons	Parcels, administrative boundaries	tens of megabytes
Raster imagery	Satellite scenes	hundreds of MB to terabytes
Elevation models	DEM tiles	gigabytes per region

Modules 5–7 cover data formats and sources in depth.

2. Software

Software is what operates on the data. The core capability is the spatial engine — the set of functions that understand geometry. A modern spatial engine offers:

Geometric primitives (intersections, buffers, distances).
Coordinate transformations (reprojecting between CRSs).
Raster algebra (cell-by-cell arithmetic).
Network routing (shortest paths, isochrones).
Spatial indexing (R-trees and similar, so queries scale).

Popular choices:

Desktop: QGIS (open source), ArcGIS Pro (commercial), Atlas (browser-based).
Server / database: PostGIS, Oracle Spatial, SQL Server Spatial.
Libraries: GDAL/OGR, GEOS, PROJ, Shapely, GeoPandas, Turf.js.
Web: MapLibre GL, Leaflet, OpenLayers, Deck.gl.

No single tool covers everything. A production GIS stack usually combines three or four.

3. Hardware

Hardware used to be the bottleneck. Early GIS required expensive UNIX workstations with dedicated graphics cards to render a road network smoothly. Today, a mid-range laptop can handle millions of vector features interactively, and cloud compute is elastic.

Hardware still matters in specific scenarios:

Large raster analysis — a single Sentinel-2 tile is ~600 MB. Analysing an entire country in one run can saturate RAM and disk I/O.
LiDAR point clouds — billions of points per project; processing them benefits from NVMe storage and multi-core CPUs.
Field data collection — ruggedised tablets, GNSS receivers, drones, depth sensors — all hardware choices that affect data quality.
GPU-accelerated rendering — 3D visualisation and deck.gl-style aggregations leverage GPUs.

For most analysts today, the hardware decision is: laptop plus cloud when the laptop falls short.

4. Methods (procedures)

Methods are the workflows and analytical recipes applied to data. A method is reproducible: given the same inputs, it yields the same output. Methods include:

Data preparation — cleaning, reprojecting, joining, filtering.
Analysis — buffers, overlays, interpolations, classifications.
Modelling — site suitability, flood inundation, accessibility.
Validation — accuracy assessment, cross-validation.
Reporting — map layouts, dashboards, exported tables.

The shift towards analysis as code — Jupyter notebooks, QGIS Processing scripts, Python pipelines — has made methods easier to share and audit. A GIS without documented methods is a collection of one-off map files; a GIS with them is a science.

5. People

The human component is the most overlooked and the most important. Roles include:

GIS analyst — performs day-to-day analysis and map production.
GIS developer — builds custom tools, integrates GIS into applications, automates pipelines.
Spatial data scientist — applies statistical and ML methods to spatial data.
Cartographer — designs the maps that communicate findings.
Data engineer (geo) — builds data ingestion, storage, and serving infrastructure.
GIS administrator — manages servers, permissions, and user accounts.
Decision maker / domain expert — interprets results, frames questions.

Large organisations may have all of these; a small environmental consultancy might have one person wearing every hat. Either way, a GIS without humans to interpret its outputs is a very expensive chart generator.

The sixth (optional) component: network

Some textbooks add network as a sixth component — the connective tissue that moves data between storage, compute, and users. In 1995 this was an afterthought; in 2025 it's the assumption. Whether a team collaborates over the internet, integrates cloud-hosted data, or serves web maps to millions, the network underpins everything. If it's unreliable, a modern GIS becomes unusable.

How the balance shifts

The relative importance of the components has changed over decades:

Era	Dominant cost
1970s	Hardware (mainframes, plotters)
1980s–90s	Software licensing
2000s	Data acquisition
2010s–20s	Methods and people (cloud + open data made compute and data commodities)

Open data portals, free satellite imagery, and open-source software have flipped the equation. If you're starting today, the expensive parts are hiring skilled people and building rigorous methods — not buying hardware or data.

A mental model for evaluation

When you assess a GIS project — your own or someone else's — ask five questions in order:

Data — is it authoritative, current, and at the right scale?
Software — does it support the required operations and formats?
Hardware / network — is performance adequate at realistic data volumes?
Methods — are workflows documented, reproducible, and validated?
People — do the right humans have access, training, and decision authority?

A failure in any layer cascades. A beautiful dashboard built on outdated data misleads. A rigorous method run on a single laptop can't scale. A world-class analyst without governance authority can't deliver change.

Self-check exercises

1. List the five classical components of a GIS.

Data, software, hardware, methods (procedures), and people. A sixth — network — is sometimes added to reflect the connectivity dependencies of modern deployments.

2. Why has the relative cost of hardware decreased over time?

Commodity hardware and elastic cloud compute mean analysts can rent compute by the hour instead of buying expensive workstations. Data transfer and specialised formats (cloud-native rasters, vector tiles) further reduce the need for local beefy machines.

3. Which component is most often overlooked in a failing GIS project?

People, followed by methods. Projects frequently over-invest in software and under-invest in training, governance, and documented workflows. A capable team with modest tools almost always outperforms an ambitious tool stack with no one to maintain it.

Summary

A GIS is a socio-technical system with five (or six) interlocking components.
Data and methods are today's expensive components; hardware and licensing costs have shrunk.
Evaluate projects by walking through each layer — weakness anywhere limits the whole.
People and governance matter at least as much as the tech stack.