4 Spatial Data Models and Quality

How spatial data are organized

Geospatial data represent both where something is and what it is. A data model determines how locations and their associated information are organized. The two foundational models are vector and raster, and each is suited to different kinds of geographic information.

Vector features and attributes

Vector data represent distinct geographic features with geometric shapes. Points mark locations, such as wells or bus stops; lines represent linear features, such as roads or streams; and polygons outline areas, such as parcels, lakes, or administrative boundaries.

A feature's geometry is linked to Attributes, which are descriptive values commonly stored in a table. For example, a road layer might contain one line feature per road segment, with fields for road name, surface type, and speed limit. Attributes can be text, numbers, dates, or other supported types. A field's meaning, units, and permitted values should be clear; for example, specifying the units makes a field named speed_limit more useful.

Raster grids and cell size

Raster data divide a region into a regular grid of cells, also called pixels. Each cell stores one or more values. A land-cover raster might store a category such as forest or water in each cell, while an elevation raster might store a height. Raster data are common for continuous surfaces and imagery, while vector data are often convenient for distinct, bounded features.

Cell size affects the level of spatial detail a raster can represent. Smaller cells usually capture finer detail but require more storage. Cell size does not, by itself, guarantee that the stored values are accurate.

Formats and dataset components

A spatial dataset commonly combines geometry or grid values with attributes, a coordinate reference system, and information about how the data are organized. A data model describes this organization; a file format is a particular way to store or exchange it. Formats vary in what they support, so a file extension alone does not tell you everything about a dataset.

Examples include GeoPackage and GeoJSON for vector data, Shapefile for vector features and associated attribute tables, and GeoTIFF for georeferenced raster data. Formats have different capabilities and constraints. A Shapefile layer supports one geometry type, and its data may be distributed across several related files. GeoTIFF can store raster values along with georeferencing information. Before using or sharing a dataset, check its format, attribute definitions, coordinate reference system, and any accompanying files.

Assessing data quality

Data quality describes characteristics that matter when deciding whether data are fit for a particular use. Quality is not simply good or bad: a dataset suitable for regional planning may be too coarse or imprecise for locating a property boundary.

Common quality dimensions include:

• Positional accuracy: how closely mapped locations match their real-world positions.

• Attribute accuracy: whether recorded descriptions and measurements are correct.

• Completeness: whether expected features or records are missing, or unwanted ones are included.

• Logical consistency: whether the data follow their stated rules, such as whether polygon boundaries connect as expected.

• Lineage: the source data and processing steps used to create the dataset.

Quality checks should reflect the intended use. For a road network, checks might include whether road locations align with reliable reference data, whether names and classifications are correct, and whether there are gaps or invalid connections. A description of limitations is as important as a quality score: users need to know what was checked, how it was checked, and what results were found.

Metadata for discovery and use

Metadata is information that describes a dataset and helps people find it, understand it, evaluate its suitability, and use it correctly. Geospatial metadata typically records the dataset's purpose and contents, geographic and time coverage, data owner or contact, creation date, coordinate reference system, attribute definitions, distribution details, quality information, and lineage.

Standards such as those maintained by the Federal Geographic Data Committee organize metadata descriptions so they can be interpreted and shared consistently. Recording metadata while data are created helps preserve details that may otherwise be forgotten. To judge whether a dataset suits a task, examine its accuracy, completeness, consistency, lineage, and metadata—not just its appearance on a map.