Data Integration and Visualization

Question 1

Visualization

Answer 1

Graphs:
Lines for trends
Dots for focus on individual values
Bars for comparison
Follow conventions
Label your axes
Start at axes 0, and make proportional the data
No pie charts

Question 2

Data Integration: 4 principles

Answer 2

• Give all things a name
• Make a graph of relations between things
• Make sure all names are URIs (are addresses on the web): Allows integration of data from different owners at different locations
• Add semantics (= predictable inference) —> explicit & formal semantics: Assign types to things; Assign types to relations.

Question 3

How to data integrate?

Answer 3

RDF schema
OWL
constructs:
symmetric: a property always holds in both directions
asymmetric: a property never holds in both directions
irreflexive:
functional: a property has ONLY ONE value for any particular instance
Inverse functional: a value for the property uniquely identifies an instance
inverse: a property is always inverse of another property
transitive:
cardinality,
equality,
inequality,
disjoint: two properties never co-occure
equivalent: two properties always co occure
complement.

Question 4

5-star data (Tim Berners Lee)

Answer 4

Tim Berners Lee —> 5-star data
1-star: Data published on web
2-star: Data published as structured data
3-star: Data has an open format
4-star: Data has a linked-data format
5-star: Data is connected to other data