Chapter 2 lecture

I'm going to rewrite my notes from class here to get things to sink in.

Data Models
On Monday we had a lecture on data models! These exciting things include objects in a spatial database plus the relationship among these objects. We work basically with 2 types of data models: vector and raster. Also with Triangulated Irregular Networks (TIN) but not as much.

Vector data models use coordinates (x,y) the "where" and attribute data (descriptive info about features) the "what" to define discrete objects- things with definite starting and stopping points. The three types of vector objects are points, lines, and polygons. Points might represent a light pole, gas wells, a tree, survey points. Attribute data are associated with each point and describe the non-spatial characteristics of the points. Lines might represent roads or rivers. They are usually represented as sets of coordinates. Attributes can be associated with the whole line, line segments, or nodes (starting and stopping points of line). Polygons can represent parks, buildings, or counties and are created by a set of connected lines. Attribute data can include perimeter, land cover type, or county name. (Representation depends on many things, including detail, accuracy, and use of data.) Vector data models are complex data structure defined by many x,y coordinates.

Vector data can be topological, which definitely doesn't mean topographic. Topology refers to the spatial relationships among features. It is a branch of geometry that I've not heard of before today. It makes up the rules that allow us to work in GIS. Words such as "connected, adjacent, and contained" are associated with topology. Computers need topology because they don't have experiences to make connections nor do they have nifty intuition so they need rules.

On to raster data models....they are simple data structure, work with jpeg formatting (digital photographs), involve continuous spatial features such as precipitation, elevation, and slope. They define the world in a regular grid pattern with a regular set of cells. Each cell has the same dimensions. "The phenomena or entities of interest are represented by attribute values associated with each cell location" (Bolstad 40). Usually each cell has one value; it is a minimum mapping unit and it cannot map anything smaller. If water and land exist in the same cell, the computer will decide which value to assign to the cell. The book explains that there is a trade off with raster data sets dealing with spatial data and data volume. Talks more on 41. Resolution and pixels come into play the raster data. Downtown Asheville is mapped at 6" resolution which apparently is very good.

Most data may be represented in either raster or vector models and can be converted between the two.

Triangulated Irregular Networks are good for representing surfaces, like elevation. Actually uses vector data generated from coordinates. Starts with points, then connects them to form a connected network of triangles. More on 50 and 51.

Data Formats
We use specific data formats to store and display information. Examples are .doc (text, images) .txt (text only, easier and quicker to open, don't need MS word). All have advantages and disadvantages. Spatial data formats are no different:

A geodatabase is a central storage location for all GIS data and relationships among them.
1. personal database (.mdb)
2. file based database (.gdb) *preferred*
3. scalable geodatebases (3)

My notes get fuzzy here, I should ask some clarification questions before I continue.