Spatial Analysis with Raster Data

Question 1

What is continuous data?

Answer 1

It is a single variable or piece of information that is defined t every location.

Question 2

How can you represent a discrete world?

Answer 2

You can represent a discrete world which is where data is discrete an object is a object it has an outline and an extent e.g. a polygon, line or a point like a bus stop location. It has a location and a start and end.

Question 3

How is continuous data defined?

Answer 3

By a square grid and it is a resolution, and by resolution we mean the size of each one of those grid cells that represent the real world.

Question 4

What spatial analysis can we do with Raster data?

Answer 4

1) Raster data structure
2) Raster Algebra
2) Comparisons and Normalisation, e.g. Reclassify and Normalise range
3) Multi-criteria analysis

Question 5

What is the structure of Raster data?

Answer 5

Digital camera images are a grid of squares called cells or pixels which have values. In Raster Data we can only store a number. Each cell or pixel can only store one number.

Question 6

How do we create Raster surfaces?

Answer 6

From vector feature data, we can use Interpolation from sample points and calculate Raster’s of distance and Raster’s of density.

Question 7

Give and example we analyse existing raster surfaces.

Answer 7

Slope of land.

Question 8

What is raster maths?

Answer 8

You can have two rasters for example and you apply a function to them to get and output where bidmass is important.

Question 9

What does the raster calculator do?

Answer 9

Allows expression (functions) to be run on one or more raster surfaces.

Question 10

What does re-sampling do?

Answer 10

Re-sampling changes the spatial scale of the raster- information may be lost.

Question 11

What does re-classifying do?

Answer 11

Reclassifying changes the VALUES in a raster- scale stays the same.

Question 12

What is normalisation?

Answer 12

It is normalising data into a standard range. E.g. 23-455 into range 0-1. Normalisation is the process of scaling raster values to a common range.

Question 13

What are the two way to normalise?

Answer 13

• Can normalise in standard fashion
Largest value= 1
• Or can carry out a reverse normalisation
Smallest value= 1

Question 14

What is the normalisation equation?

Answer 14

Value(ij)= v(ij)-min(vij)/max(Vig)-min(vij)

where ig is the current raster cell.

Question 15

What is the reverse normalisation equation?

Answer 15

Value(ij)= max(Vig)-V(ig)/max(Vig)-min(Vig)

where ig is the current raster cell.

Question 16

What does MCE stand for?

Answer 16

Multi-criteria evaluation

Question 17

How can MCE be carried out?

Answer 17

Using raster algebra once various data sets have been normalised.

Question 18

Example of using MCE: We wish to build a new (set) of metro stations to extend the current network:

1) Must be an area of flat land
2) Must be an area where there is demand for public transport services
3) Must extend the current network geographically.

Answer 18

1) Flat land can be derived from a DEM using SLOPE
The flattest areas should be assigned the highest value.
2) Public transport demand can be derived from the numbers of bus stops (Surrogate data)
The more dense the bus stops the higher the value assigned.
3) Distance from existing Metro stations can be derived from the metro stations.
The furthest distance is assigned the highest value
4) Weights can be added to each criteria.
Public transport demand is more important than slope (we can always dig).

Question 19

What must the weights sum up to?

Answer 19

The weights must sum to 1.

Question 20

Define raster data.

Answer 20

Raster data is a matrix of values representing a single variable continuously and Raster data can be used in complex analyses.

Question 21

What is MCE?

Answer 21

MCE is a process used to evaluate multiple inputs

Used for site selection analysis.