Lecture 2 - Spatial Data Flashcards
Tobler’s Law
Everything is related to everything, but near things are more related than distant things
Also called: spatial dependence
What is spatial data?
Also known as ‘geographic info’ or ‘geospatial data’
- data or info identifying features & boundaries on earth (abstraction of real-world)
- natural or constructed features
- stored as coordinates, topology, or surfaces
- can be mapped
- can be analyzed in GIS
- then we can start to understand the phenomena
Why is spatial data important?
Space is a fundamental component in our everyday life (together with time)
We can use it for analysis and this can help with decision making or understanding a phenomena
Examples:
- how do we choose a house to live?
- how can a disease spread?
- where should I start my new business?
Types of spatial data (4)
Data can be:
- events (incident points)
- continuous (surface of temperature or elevation)
- zonal (polygons like regions or meshblocks)
- spatial interaction (transfer of goods, data, information, traffic)
SPATIAL DATA ANALYSIS INFO
Data –> information
- beyond just mapping
- transformation, manipulation, and analytical methods
Collection of tools for investigating
- spatial patterns of values & places (need location and attributes)
- variations of spatial phenomena based on their associations
Location-dependent
- move object location and result will change
- median centre, clusters, spatial autocorrelation
SPATIAL DATA AIMS and PROCESS
Aims:
- identifying, describing and understanding the PATTERN and PROCESS
- infer the process by first observing the pattern
Process:
- Exporatory Spatail Data Analysis (EDSA) –> find interesting patterns
- Visualisation –> show interesting patterns
- Spatial modelling, regression –> explain interesting patterns
What is spatial analysis?
The analysis of data for a process operating in space, where methods are sought to describe or explain the behaviour of this process and its relationship to other spatial phenomena
Objective: understand the spatial arrangement of variables, detect patterns, examine relationships of variables
How is spatial data represented?
Represented by 5 entity types: vector (point, line, polyline, polygon) or raster
Attributes are spatial and non-spatial (mostly interested in non-spatial but want to study with their location)
Geocoding
Transforming location descriptors (postal address, place name, Zip code, land parcel ID, etc) into single (x,y) coordinate location in order to map the location as a point
Geocoding issues
- Place name ambiguity (some places have the same name)
- Incorrect/incomplete addresses
- Abbreviation handling
- Names might be local
- Incorrect spelling
- Single field addresses
- Addresses for open space (or large areas)
- Non-existent addresses
- Verifying (if you can’t physically check the location there could be an issue - the conversion is not perfectly accurate)
Spatial Analysis Steps (6)
- problem formulation
- data gathering
- exploratory spatial data analysis
- hypothesis formulation
- modelling & testing
- consulting & review
Human Vs Natural Data
Spatial data can be both natural or human
Natural domain: Images (satellite or drone images of earth’s surface), environment (topography, soil, temperature, hydrology, geology).
Human domain: Cadastral (parcels, streets, land use), census (demoographics, economics)
Infrastructure (human or natural): transportation, energy, water, waste, communications, monitoring
Spatial analysis parameters (4)
Distance:
- Euclidean
- Network
- Travel time
Adjacency:
- Shared boundary
- Within X distance
Interaction (Distance + Adjacency):
- Tobler’s 1st Law
- IDW
Neighbourhood:
- set of entities adjacent to an entity
- space region associated with an entity (defined by distance)
When to use spatial analysis
Is there a reason to suspect differences across space?
- differences in phenomena (variables)
- differences in relationships between phenomena (covariance)
Does data have any spatial reference?
