Thinking spatially

1
Thinking Spatially
JWAN M. ALDOSKI Geospatial Information Science
Research Center (GISRC), Faculty of Engineering,
Universiti Putra Malaysia, 43400 UPM Serdang,
Selangor Darul Ehsan. Malaysia.
2

• Spatial data exploration - How to represent our
  world and environment digitally, graphically and
  numerically
• Important considerations
• Scale
• Aspatial/spatial data
• Discrete/continuous data
• Measurement
• Location
• Spatial relationships

3
Effect of scale on Spatial dimension A house on
a close-up aerial view appears to have length
and width, but as we pull back its length and
width dimensions disappear, leaving as with house
as a point
4
Spatial Elements

• Fundamental classification of real world
  phenomena and objects represented in a GIS
• 4 basic types points , lines , areas (polygons),
  and surfaces
• Points, lines and areas can be represented by
  symbols
• Surfaces are represented by point elevations
• All data are explicitly spatial
• Attribute - Aspatial data

5
Spatial Elements
6
Surface features

• Continuous objects or phenomena - surface
  composed of infinite number of Z values -
  sampling is used to represent surface in GIS
• Stored as a grid of cells in raster GIS
• Stored as a Triangulated Irregular Network (TIN)
  in vector GIS (special kind of vector data)
• Physical surfaces - e.g., topography
• Abstract surfaces - concentration of pullutants
  in a body of water
• 2.5D vs. 3D

7
(No Transcript)
8
Discrete and Continuous data

• Continuous data objects which have no definite
  boundary, generally no "empty" space and assumed
  to have three dimensions X,Y and Z e.g.
  elevation, temperature, and rainfall. The data is
  represented as surface in GIS
• Discrete data objects which occupy a specific
  location in space at a given point in time e.g.
  road, river, and lot.and represented as point,
  line, or area feature in a GIS

9
Continuous vs. Discrete
10
Measurement Levels

• Attributes - information which describes the
  spatial entities, usually stored in a database
  linked to the map features. Levels of measurement
  used to describe data (spatial and aspatial)
• Nominal scale data
• nominal name , allows us to make
  differentiation between objects, does not allow
  us to rank objects, does not permit quantitative
  comparison e.g., land cover types (forest,
  water, roads, urban, etc.)

11
Measurement Levels-continue

• Ordinal scale data
• ordinal order , allows us to rank
  observations, does not permit
  quantitative comparisons, e.g.,
  good, fair, poor large, medium, and small
• Interval scale data
• numbers are assigned to observations/entities,
  permits quantitative comparison,
  e.g., soil temp. measured at Farenheit scale,
  does not permit ratio level comparisons e.g.,
  soil at 50 deg. is not twice as warm as a soil of
  25 deg., why? 0 on the scale in interval
  measurement is arbitrarily defined (does not
  represent the absence of quality
  being measured)

12
Measurement Levels-continue

• Ratio scale data
• Numeric scale is defined by an absolute 0
  (e.g., Kelvin temp. scale), e.g., monetary scale
  permits quantitative comparison as
  well as computation of ratios

13
Measurement Levels-continue

14
Populations and Sampling Schemes

• Sampling - looking at a subset of individuals in
  a population to make some inferences about the
  entire population
• Target population - collection of all objects or
  phenomena of which we wish to know something
• Sampling area - spatial extent of area being
  sampled
• Sampling frame - target population sampling
  area

15
Types of sampling strategies

• Random - all entities have equal probability of
  being selected
• Systematic - selection of entities based upon
  some systematic design e.g., every 10th tree in
  a transect, soil temperature collected every 100
  feet
• Stratified - dividing the population into spatial
  subsets or thematic subsets before sampling
  e.g., X number of samples are to be taken
  from each of 4 plots
• Homogeneous - opposite of stratified, all objects
  in population are considered part of one
  homogeneous group, no separation or
  stratification is used

16
Making inferences from spatial data

• Sampling leaves gaps in our knowledge of
  unsampled areas
• Need to predict the missing points

17
Spatial data model

• Vector model
• Raster model

as geometric objects points, lines,
polygons
as image files composed of grid-cells
(pixels)
18
PLEASE Thinking Spatially

• AND THANK YOU