Isabel Meirelles,
Associate Professor of Graphic Design at Northeastern University
- Hierarchical Structures are ordered sets where elements and/or subsets are organized in a given relationship to one another, both among themselves and within the whole
- Nested or stacked structures
- We are dealing with an abstract concept
- Map can be hierarchical: country, state, province, …
- Best representation to depict the hierarchy is to find the best conceptual metaphor, not the visual metaphor
- Abstract data: showing a tree structure
- Physical data: non-overlapping boxes
- Metaphor (of the tree)
- Schema: the tree is a part-whole configuration for structuring a target concept
- Structural elements: a whole, parts and a configuration. Because the parts can exist without constituting a whole…
- Trees
- Trees and “arbores had a solid authoritative base in the Bible. The Book of Genesis speaks of the tree of life planted by God in the Garden…”
- In early work, people used the roots of the tree as part of the visualization
- We have also seen uses of the vine
- Using the human body diagram, which looks like a tree,
- Also the house as a metaphor as well as the book
- IBM Many Eyes
- Isabel has written a book on these topics
- Visualizations
- Ernst Haeckel “Paleontological Tree of Vertebrates”
- Henrich … polar system
- Dual graphs, where one is the tree, and another is a nested circular treemap, which shows the clusters that are invisible in the tree
- Recent techniques
- 3D visualizations: George Robertsom, Snapshot of the “Cone Tree” visualization technique, 1991
- Johnson, Snapshot of the “TreeViz” interface that uses treemap to represent files in a computer
- Munzner, snapshot of the “3-D Hyperbolic Trees”, 1998
- Relational Structures organize data for which relationships are key to the system being visualized
- Food networks, metabolic networks, power structure (World Finance Corporations), “They Rule” 2004 by Josh On, complexity of American military strategy, small partial map of the internet using data from the Opte Project by Matt Britt, Subway map
- Origins of Network Science
- Euler puzzle: traverse the bridges only once, travel to all places
- Beginning of graph mathematics
- lands were nodes, bridges were edges
- Figured out it was not possible
- “A network is a simplified representation that reduces a system to an abstract structure capturing only the basics of connection patterns and little else. Vertices and edges in a network can be labeled with additional information…” – Mark Newman
- Says math calls node and edge as “site and bond” or “actor and tie”
- Definitions:
- Ranking: ordered sequence of friends
- Type: friend, relative, co-author
- Layout types
- linear
- force-directed
- circular
- Bundling chords going to same direction
- http://well-formed-eigenvector.com (I think)
- community structure
- geography based
- sankey type
- Fineo by Density Design Research Lab, Italy 2010
- Over the decades, how states have shifted, by Mike Bostock
- force-directed centered on node
- polar
- radial community structure
- Google+ ripples: how messages have been disseminated on Google+
- matrix
- Takeaways
- Questions influence visualizations (descriptive vs argumentative)
- It might be about the process rather than the output
- Solving a design problem requires finding the right representation
- Visualizations largely depend on the task and audience
- Perceptual and cognitive constraints (going against the metaphor)
- Visualizations are human-centered
- Book: Design for Information
- Leonardo
- Workshop