Visualisation and Colour

Question 1

Tools for spatial visualisation

Answer 1

Scale, projection and symbolisation

Question 2

Basics of a well presented map

Answer 2

Short, concise, clear title
Scale bar
Clean legend: no abbreviations, clear title
North arrow
Map balance: not too much white space
Readable
Does my symbology support my story

Question 3

Symbolisation

Answer 3

what symbols we use to communicate attribute values associated with geographic locations → points, lines and polygons

Question 4

Basics of design

Answer 4

Meaningful colour/shading
Visual contrast: to differentiate land/area, symbology
Legibility: symbols and fonts readability
Cam change colour, font, orientation
Figure-ground organisation; what are you looking at
Hierarchical organisation: bringing the most important thing to the top
Balance: limit white space, make

Question 5

Choropleth Maps

Answer 5

Colour-coded polygon maps

Represent numeric values (population, number of cases, % of people vaccinated)

Question 6

Colour: 3D concept

Answer 6

Hue: red, green, blue
Easiest to differentiate
Different hue to represent different categories of data
Lightness: value
Easiest to order
Important dimension for representing ordinal/interval/ratio data
Saturation: chroma, vividness, brightness
Viewers can order but not as well as brightness
Modified sparingly in most maps

Question 7

Why use colour?

Answer 7

To show order, to show different

Question 8

Qualitative colour scheme

Answer 8

Used to symbolise data having no inherent order (categorical data)
Different hues usually to differentiate between categorical values - when possible use commonly associated for land features (water, vegetation)
**avoid pink and turquoise as it is unprofessional
Be careful in choice of hue if a cultural association exists

Question 9

Sequential colour scheme

Answer 9

To highlight ordered data such as income, temperature, elevation, infection rates
Ranges from a light colour (low attribute values) to a dark colour (high attribute values)
Typically a single hue but may include 2 hues

Question 10

Divergent Colour Scheme

Answer 10

Apply to ordered data - but an implied central value about which all values are compared
Typically 2 hues - one for each side of the central value
Each hue’s lightness/saturation value is then adjusted symmetrically about the central value

Question 11

Colours: common cultural and scientific associations

Answer 11

Disease cases in multi-colour scheme
red/orange for higher number
green/blue for low counts
Yellow in between
FLIP for accessibility
Blue and green for wet and browns/yellow for dry
No data should be grey and zero cases white

Question 12

Cyan

Answer 12

the default colour for selection - make sure it is in the legend or turn off selection

Question 13

Which classification to use

Answer 13

Natural breaks/Jenks: data aren’t evenly distributed or have gaps
Quantiles: want to compare areas of about the same size and values are evenly distributed
Equal interval: even distribution
Standard deviation: want to see areas above or below the average (normal distribution)
Geometric interval: need a compromise method between equal interval, natural breaks And quantile

Question 14

Don’ts

Answer 14

DO NOT USE DIAMOND FOR POINTS - squares, triangles and circles work well
**do not mix combos of colour/shading and cross hatching to represent data classes → makes it difficult to determine sequential valueVegetation - generally do not want borders

Question 15

Categories

Answer 15

usually want 5-7 categories
Too few breaks = too much heterogeneity
Too many - lose pattern
Zero inflated data - example of when to exclude data

Question 16

Choropleth map choices

Answer 16

Shading
Hue
Dots patterns
Dots density

Question 17

Dot density map

Answer 17

A random pattern of dots placed in each region with the number of dots proportional to the attribute value associated with that region - advocated for business uses (number of customers), maintains some confidentiality
Need a disclaimer to explain dots are randomly distribute
Dot density maps DANGEROUS
Dots do not represent real locations but the viewer cannot help but interpret them that way - and reproduced map by lose accompanying text
Dot density maps too easy to misinterpret to use with sensitive data such as disease incidence/prevalence

Question 18

Classifying data

Answer 18

process of placing data into groups/classes that have a similar characteristic/value
Jenks tries to minimise differences within classes and look for clustering of values → natural breaks in the data
Create manual breaks - good when there are industry standards, epidemiological alarm cut offs
QGIS: pretty breaks = breaks the values into classes that are easily understood by non-statisticians