Geographic Representation and Data

Geographic Representation

• Fundamental Problem: Simplifying geographic data for map representation (raster and vector). Reducing the Earth's complexity to a map projection.
• Representing Data: Creating digital maps of the Earth's surface for visualization beyond our immediate experience.

Geographic Data

• Geographic Datum: Links a location and its descriptive properties (attributes) over time.
• Attributes: Descriptive elements of geographic objects. Example: "arable land" can be categorized as "tilled" or "fallow."
• Attribute Classification:
  • Nominal: Unordered categories (e.g., land cover types, species).
  • Ordinal: Ordered categories with non-uniform intervals (e.g., soil quality: low, medium, high; study grades A-D).
  • Interval: Numerical data with meaningful intervals but no absolute zero (e.g., temperature in Celsius).
  • Ratio: Numerical data with meaningful intervals and an absolute zero point (e.g., distance, concentrations).
  • Cyclic: Geographic attributes with repeating values (e.g., compass directions). Hierarchical structure; each level builds on the previous.

Solving the Fundamental Problem: Simplification

• Spatial Averaging: Assigning a constant value to pixel/tessellation areas (regardless of size). The impact of pixel size (dpi) dictates detail : Smaller pixels (high resolution) vs larger pixels (general overview).
  • Raster Data:
    • Pixel: Constant value; significant space consumption due to large ocean areas having the same value.
    • Tessellation: Mosaicing the Earth's surface; large areas of constant values.
• Discrete Objects: Objects occupy space; clear boundaries.
  • Identifying Objects: Based on dimensionality (points, lines, areas)
  • Limitations: Difficulty in representing truly 3D objects and continuous natural features (e.g., rivers, mountains).
• Continuous Fields: Variables measured across the entire surface; values vary continuously. Examples include population density, traffic density, and elevation.
  • Types of Variables: Nominal, ordinal, interval, ratio, cyclic.
    • Scalar fields: One variable per point
    • Vector fields: Two variables(e.g. magnitude and direction) per point

Raster and Vector Data

• Raster: Finite set of continuous surface info; used for elevation, soil types, etc. Derived from remote sensing images. Values are assigned by central point or largest share (largest share more common).
• Vector: Finite set of object information.
  • Representing Curves: Using connected points(vertices).
  • Polygons: Closed shapes; representing areas (e.g., parks, lakes)
  • Polylines: Representing curves using connected straight line segments. Efficiency in space compared to raster depictions of curves.
  • Dimensionality: Points (0D), lines (1D), polylines (2D), polygons (2D)
    • Points, lines, polylines, and polygons, each with identities and coordinates.
• Raster vs. Vector:
  • Precision: Vector is precise, raster is approximate(suited to statistical randomness)
  • Storage: Vector more space-efficient with large details; raster good for statistical analysis.
• Paper Maps vs. Digital Maps: Paper maps have scales; digital maps can be zoomed in.

Generalization

• Simplification: Reducing complexity; various methods, such as "weeding" (removing points on vector lines).

Sampling and Uncertainty

• Sampling: Geographic data often sampled rather than complete. Uncertainty exists between samples.
• Statistical Tools: Measures and minimizes errors. Statistical standard errors relevant.
• Autocorrelation: Similar values in geographically proximate areas; important for predicting missing data.

Autocorrelation

• Importance: Modelling and interpolation of missing data, along with relationships and distances between readings.
• Applications: Modeling and data interpolation in geographic analysis. Improves data accuracy by considering positional relationships and consistency.