Geographic information system vs Science

(epic battle)

What is geographic science?

A fundamental field of study which examines the representation, storage, analysis and visualization of geographic information

• Basic research field that seeks to redefine geographic information/concepts and its use in the context of GIS

• Real world things are entities (geographic themes) conceptualization of spatial entities governed by how we percieve or interpret that entitty and its intended applicatoin in GIS

• Object: discrete entities with boundaries and locatoin explicitly defined; modeling the human/urban landscape dominates this view
• Field: A collection of spatial distributinos that vary continuously acros space; modeling the natural environment dominates this view

• Both views can be used to represent the same entity, they are not mutually exclusive in representing the geograpic phenomenon.
• ex. Lakes could be represented as discrete objects with defined boundaries, e.g. shorelines, named etc.
• Lakes could also be represented as a continuous distribution of the degree of "lakeness" i.e. dry, seasonally wet, etc.
• depends on the users percetion of that phenomen

• representitive form of the conceptual view
• real world observations modeled by spatial geometry (coordinates;georeferenced) and attributes (data and text describing the characteristics of the features)

• Vector: represents points, lines, polygons, locatin explicitly defined by pairs of coordinates (points is the basic building block

• Raster: cells, location implicit by the size/area of the cell and the cell layout (cell is the basic building block)

Vector: object view, best for discrete data with well defined boundaries

Raster: Field view, best for continuous data

however, the same view can be converted and modeled in both data models

• Are georeferenced to either a geographic of projected coordinate system
• are defined by a particular mad scale (an original sacle of capture) i.e. spatial resolution
• are defined by a characteristic scale of use or application (governed by its map scale)
• are smaller than the reality they represent; limits the amt of detail
• involve some level of generalization of simplification

More key features in spatial data

are flawed as a function of our conceptualizatino of an entity, our definiton of an entity, its intended purpose in GIS, and or by the limits in technology for "capturing" that entity as data e.g. satellite resolution

Spatial data structure (physical level)

(Vector)

Vector dimensionality and properties

• zero dimensional (no length or area; point, node)
• one D ( no lenght, no area; line segment, arc)
• 2D (lenght, area; polygon)
• The chosen dimenionality of geograpic represntation is influenced by map scale ( river is a piont at 1:5,000,000: river is a line at 1:500,000; river is aplygon at 1:5,000)

Spaghetti Vector Model

• unstructured method of representing simple points, lines and polygons
• No spatial relationships (topoloty) are explicitly defined or encoded with data
• simplest of vector structure
• common data structure: shapefile or CAD file

• The spatial relationships between features that do not change; a set of rules and behaviors that model how points, lines, and polygons share geometry
• topological associations: connectivity, contiguity, containment

• Uses related tables to explicitly define and encode spatial relationships (topology)
• tables store information on Directionality, left and right polygons, from (srart) and to (end) nodes
• complex vector structure; most common approach is the arc-node model
• common data structure: Arcinfo coverage

• Based upon an abstractino of features into database objects stored in a relational database management system
• objects can intergrate feature "behavior" through spatial (topological) relationship classes. and attribute domains; these rules and relationships are stored in tables
• common data structure: geodatabase feature class and feature dataset

• attributes share a one to one relationship b/t spatial features and their corresponding record in the value attribute table
• why is this significant?
• There is always one record in the attribute table that corresponds to a feature on the map allowing for multiple records in the table to have the same "uniqe" value. In addition if a record in the table or a feature ihn the map is deleted, so is its corresponding feature or record, respectively

TIN (triangulated irregular network)

• Vector structure used to model surfaces, i.e. continuous data, like elevation
• based on irregularly spaced points with associated coordinates in three dimensions (x,y, and z) connected by edges (lines) that form a network of non-overlapping triangles

Raster

How is it represented?

cells, each cell can have only one value, stored as a number

• raster resolution is the amt of detail the raster dataset contains, is a functio of the cell size/area at the original tate of capture
• resolution (basedon cell size) cannot be increased by simply "resampling" to a smaller cell size

The mixed pixel problem associated with cell value assignment

(raster)

• cells are  characterized by size and occupy area, thus assigning a single attribute value to a cell coverng an area with hetergeneous (multiple) classes, e.g. water, agriculture, forest, in the real world is a problem
• Affects the overall accuracy of a dataset and its attributes, the appropriateness of a dataset for particular application and its use at a particular scale

• cell attributes are stored in numbers, either integer(for discrete rasters) of floating point (for continuous, statistical surfaces)
• each cell must have an attribute value, even though it may be "missing" or "null"; usually defined as a NoData value, e.g. -9999

Raster attributes

(theres more)

• attibutes share a many-to-one relationship b/t cells and their corresponding value in the value attribute table, why is this significant?
• W Vector data, there is always one record in the attribute table that corresponds to each feature, allowing for multiple records in the table that have the same "unique"value. In raster, the attribute table has one record for each unigue value in the grid. When makding a selection w raster data, you are selecting only one record in the table, aso opposed tomultiple records with vector data.

Database management and Query

(DATA BASE)

• a structured set of data usually organized into tables; a matrix of columns and rows
• entity (theme, table) instance, row (record); column (field)

database management sustem

(DBMS)

• Tables that are logically (virtually) associated with each other by shared fields (keys)
• keys are unique ID's or fields with common info
• A primary key is a field that uniquely identifies each record in a primary (destination) table
• A foreign key is a field from a second (source) table used to li8nk it to a primary (destination) table through its primary key

Database

(Extended)

based on a database design technique called normalization which removes repeating columns and repeating rows, i.e. duplicated data, and relates the resuldign table using keys

• the most utilized database design in GIS

• collection of data that include both spatial (geometry) and attribute stored in related database tables or files e.g. georelational
• vector data store thier attributes in a feature attribute table (FAT)
• raster data (discrete/integer) store their attributes in a value attribute table (VAT)
• exception, and why- continuous (float) raster data there is NO VAT created; since there is a potentially an infinite number of values in continuous data it is nearly impossible to create a VAT

• establishing a temporary link (connection) b/t records in two seperate tables through an attribute field (key) common to both
• Virtually appends attrubutes from ont table to another based on a primary key (destination table) and foreign key (source table)

one-to-one, many-to-one cardinality. many (or one) records in the destination table have exactly one recor foined to it from the source table

join is usually used to attach attrubutes to the attributes table of a geographic layer by appending the fields of one to the other

• one-to-many, many-to-one cardinality. one (or many) record in the destination table has many record related to it from the source table
• unlike joining tables, relating tables simply defined a relationship b/t two tables. the associated data is not appended to the layers attribute table like it is with a join. Instead, you can access teh related data with necessary.
• refer to lecture for examples and practice

• appends attrubutes fromo n layers attribute table to another layers attribute table based on the relative spatial locations of the two layers

Appends based on the spatial relationships

(spatial joins)

• intersect-> features that share a common part with others
• containment-> features that contain or suffound others
• proximity-> features within a distance of

Appends based on cardinality of the relationship

(Spatial joins)

• polygons to points: many-to-one; simple join
• points to polygons: one-to-many; summary join

refer to lecture for examples and practice

Basic strucuted query language

(SQL)

set algebra (<,>,=,<>)

boolean algebra NOT,AND,OR

be able to set queries and fill in venn diagrams

spatial query

commonly used to select featuers based on coordinate geometry...

• adjacency-> features that touch or share  a boundry with other features
• intersect- features that share a common part w others
• containment- features that contain or suffound other features
• proximity- features within a distance of