Basic concepts: Spatial and tabular data

From ILWIS
Jump to: navigation, search
Main Page Arrow.png Basic concepts Arrow.png
Spatial and tabular data
Spatial data

In general, spatial or geospatial data refers to representation of the Earth locations, their size and shape in a geographical space. Spatial data is visualized in form of points, segments (lines) and polygons (areas) using analog or digital formats. Spatial data are usually accompanied by non-spatial data, known as attribute, stored in tabular format. The appended attribute data provides additional information about each geographic feature and its characteristic. The data models used in ILWIS to create, store and visualize geospatial data are vector and raster. Both models store details on the location of entities and their attributes. The main difference between the two data models is the way they store and represent geospatial data. In the following, the data models including vector, raster, tabular and their relationship are briefly described.

Vector model

Vector data models use points and their associated X, Y and Z coordinate pairs to represent the vertices of spatial features. In ILWIS, a spatial feature is anything that can have one or more geometries, one distinct identity and a code (label) that provides the meaning of the feature. An ILWIS vector feature coverage may contains only points, lines or polygons or any combination of them.
In addition, each feature can have one or more representations. One can see these representations as evolution of the feature in a certain dimension. For example, an ice berg drifts over sea (assuming it doesn’t split) and changes its position, shape and probably its other attributes. This represents an evolution of the feature using the time attribute. The identity of the ice berg doesn’t change while its properties including its spatial properties, do. In ILWIS this considered as one feature with one or more "sub" features. Another example would be a city at different scales. At world level, its representation might be a point, at local level a complex polygon with districts; while at in between scales the polygon becomes less and less complex (depending on the scale direction). The evolution is in this case in the scale attribute. The organization of sub features is called the "stack" in ILWIS (very similar to a the same concept with raster coverages). The stack has a domain which represents its semantics ( time and scale in the examples). Note that the stack can have an ordering which dependents on the domain. Time and scale are ordered but if one would use a thematic domain there would not be an explicit ordering. Organized in a feature coverage the different entries in the stack can be seen as layers. Note that the sub in sub feature can mean both alternate and part-of. Features are, thus, described by:

  • Per stack entry a geometry ( which maybe an empty geometry, but at least one for the whole stack)
  • Stack dependent attributes
  • Stack independent attributes
Raster model

In raster data model, spatial data are organized in grid cells or pixels where position of each cell is defined by a row and column number. Pixels, a term derived from picture element, are the basic units for which information is explicitly recorded. A point in the raster data model is described by the position of a single pixel and its corresponding code (label), while a line and an area are described by a set of connected pixels having the same code.
In ILWIS, raster data stored according to the raster model is called raster coverage. Pixels in a raster coverage have the same dimension which is defined on the basis of the data accuracy and the resolution required by the user. Generally, no explicit coding of geographic coordinates is required, however, coordinates can be assigned to each pixel of a raster coverage using georeferencing. In georeferencing procedure, the parameters to transform between X, Y and Z coordinates of a location on a map and its corresponding pixel locations in the raster coverage (rows and columns) are established.
In new ILWIS, a raster coverage of single or multiple layer(s) is stored in a single file using stack concept described in vector data model. For example, multispectral images data with several spectral bands or multi temporal data of a same area are stored in one data file and the stack domain is used to organize them for visualization or data analysis.
ILWIS supports conversion of vector data to raster data or vice versa through conversion operations. Raster data are also generated by scanning hard copy images including aerial photographs or directly generated by e.g., space borne, airborne and field imaging systems. In ILWIS, most of data analyses are based on the raster data model and the output results of the most operations are in raster format as well.

Tabular data

In ILWIS tabular data is stored in table which is an arrangement of data in rows (records) and columns (attributes). Each column of the table is identified by a unique name and a domain while the records are identified via table prime key contents. This means every record shares the same set of column headers (variables). Contrary to previous versions, the tabular data stored in new ILWIS format do not requires any domain for the table. Tabular data in ILWIS can be associated with geospatial features or can be used independently as standalone table with no dependency. When a cell (intersection of a row and a column) is lacking information, ILWIS stores an undefined value represented by a ? in the cell for the missing information.

Link between spatial and tabular data

Spatial data can be linked to non-spatial (tabular) data e.g., a table. Once this relationship is established the tabular data can be referred to as attribute data or attribute table. In new ILWIS, attribute tables can be linked only to raster maps but not to vector data (feature coverage). Attribute data associated with vector data (feature coverage) is embedded and stored in the same file, even though it appears as a table in the catalog with the same name as the feature coverage. Since in new ILWS, tables do not have any domain, an attribute table can be linked to a raster coverage via one of its column if the column and the coverage have the same domain. Therefore, the number of the records in the table have be at least equal or more than the geographic features in the map. Also, it is important to keep in mind that attributes data can only be linked to raster coverages with thematic domain but not to raster maps with value domain.