iVia Notes on General Visualization, Graph Visualization or Interface:

Tools:

JUNG ****

the Java Universal Network/Graph Framework--is a software library that provides a common and extendible language for the modeling, analysis, and visualization of data that can be represented as a graph or network. The JUNG architecture is designed to support a variety of representations of entities and their relations, such as directed and undirected graphs, multi-modal graphs, graphs with parallel edges, and hypergraphs. It provides a mechanism for annotating graphs, entities, and relations with metadata. The current distribution of JUNG includes implementations of a number of algorithms from graph theory, data mining, and social network analysis, such as routines for clustering, decomposition, optimization, random graph generation, statistical analysis, and calculation of network distances, flows, and importance measures (centrality, PageRank, HITS, etc.). Also see InfoVis Toolkit. http://jung.sourceforge.net/

Tulip ****

For visualizing very large graphs using node-link diagrams... cope. with hundreds of thousands of nodes. This supports many diverse visualizations and is open source.
http://www.tulip-software.org

SPIRE and IN-SPIRE ****

http://in-spire.pnl.gov/about.html
free to fed gov projects ($500 fee) visualized crawls; supports exploration of textual data, including Boolean and .topical. queries, term gisting, and time/trend analysis tools; interactive up to 30k-60k 1 pages docs (might = bib rec + desc); UNIX and Windows flavors; pretty impressive; zoom (galaxy > topic terrain); reads ASCII or XML (we can convert). Pretty sophisticated. Beautiful. May be too experimental. Free trial. Taxonomy-Based Classification and Visualization of Text
http://www.emsl.pnl.gov/docs/cis/annual_report1999/1558b-3c.html

Starlight ****

http://starlight.pnl.gov/
U.S. Gov or sponsored org. We would qualify. Go to applications then go to Web maps whereupon we get a nice visualization of a focused crawl accompanied by the following: The image above shows a Starlight Network View representation of the "Information Visualization" Web community. This collection of Web pages was harvested using the Google search engine to retrieve the top 100 ranked pages containing the phrase "information visualization," as well as all pages that link to, or are linked to from, the pages in the query result set. The entire collection consists of approximately 1500 pages. In this view, nodes represent discrete Web pages, and edges represent hyperlink references among the pages. The pages are color-encoded according to their link class, with pages in the original result set colored green, result set inlinks colored blue, and result set outlinks colored yellow. Note that the majority of the result set pages are embedded in the clique (dodecahedra) and cluster (stellated dodecahedra) nodes. ........ Starlight is currently only available to U.S. Government and Government-sponsored organizations. If your organization meets this qualification, please contact us for licensing and availability information. Be advised that, as Starlight users, you and your colleagues will be leaving terra firma behind, and joining us out on the edge of possibility. Or something like that.

InfoVis Toolkit ****

http://www.lri.fr/~fekete/InfovisToolkit
The InfoVis Toolkit, as of version 0.6, implements eight (8) types of visualization: Scatter Plots, Time Series and Parallel Coordinates for tables; Node-Link diagrams, Icicle trees and Treemaps for trees; Adjacency Matrices and Node-Link diagrams for graphs. Naomi likes this one.

From InfoVis author Jean-Daniel Fekete <Jean-Daniel.Fekete@inria.fr> :

"I received inspiration for my data structure from LEDA graphs and the C++ BOOST Graph Library. They don't use "objects" for storing graphs but simpler containers. I just used their architecture and generalized it to tables, trees and graphs. The performance and memory footprint is much better this way than using objects for edges and vertices. Any number of attributes can be associated with vertices and edges in that way, and these attributes can be fetched very quickly and in a type-safe manner. I haven't looked at your implementation of attributes associated with edges and vertices but the trivial implementation using a vector of objects is both very slow and very memory intensive (I have made some comparisons in a tech report http://www.inria.fr/rrrt/rr-4818.html).

As for actual numbers, the Toolkit can work with million of vertices and edges but the visualization is slow above 10,000 items. This will be greatly improved with Java 1.5 when the OpenGL-based implementation of Graphics2D is available. We have implemented such a beast at the Univ. of Maryland (it is called Agile2D http://www.cs.umd.edu/hcil/agile2d/index.shtml) but Sun is currently working on a standard cross-platform implementation that will provide the speed required to visualize hundred of thousands of items at interactive speed I guess. Currently, the InfoVis Toolkit relies on executable programs to do the graph layouts (the GraphViz programs: dot, neato and twopi). Using Jung, it could be done entirely inside the toolkit with greater control. So far, the toolkit is distributed under the Q Public License, which is like the GPL but forbids forks so I can control the development until I feel it is sufficiently stable to be left alone. I would then turn it to a BSD-like license.

Don't forget that the Toolkit is a work in progress. However, it is used by some quite heavy packages and seem to stand the load so I am quite confident about its structure and extensibility. I think we should try to merge our efforts if possible to avoid fragmentation and focus on innovation and improvements instead of repetition."

Agile2D (based on InfoVis)

http://www.cs.umd.edu/hcil/agile2d/index.shtml

InfoVis CyberInfrastructure ****

http://iv.slis.indiana.edu/
This web site provides access to different software packages easing the exploration, modification, comparison, and extension of data mining and information visualization algorithms. Diverse software packages were bundled into learning modules. Links to diverse databases, compute resources, and references are provided as well. This is a compendium of resources/papers/software of value in getting an overview.

Treemap ****

http://www.cs.umd.edu/hcil/treemap Treemap is a space-constrained visualization of hierarchical structures. It is very effective in showing attributes of leaf nodes using size and color coding. Treemap enables users to compare nodes and sub-trees even at varying depth in the tree, and help them spot patterns and exceptions. Open for usage by higher ed. Not sure if it will scale to a crawl visualization.

Pajek

http://vlado.fmf.uni-lj.si/pub/networks/pajek/
large network analysis.compared with JUNG often

Walrus:

Walrus is a tool for interactively visualizing large directed graphs in three-dimensional space. By employing a fisheye-like distortion, it provides a display that simultaneously shows local detail and the global context. Walrus is a tool for interactively visualizing large directed graphs in three-dimensional space. It is technically possible to display graphs containing a million nodes or more, but visual clutter, occlusion, and other factors can diminish the effectiveness of Walrus as the number of nodes, or the degree of their connectivity, increases. Thus, in practice, Walrus is best suited to visualizing moderately sized graphs that are nearly trees. A graph with a few hundred thousand nodes and only a slightly greater number of links is likely to be comfortable to work with. http://www.caida.org/tools/visualization/walrus/

H3Viewer

http://graphics.stanford.edu/~munzner/h3/
The H3Viewer library provides layout and interactive navigation of node-link graphs in 3D hyperbolic space. The library can handle quite large graphs of over to 300,000 edges quickly and with minimal visual clutter. The hyperbolic view allows the user to see a great deal of the context around the current focus node. Walrus is the Java3D version.

GGobi

http://www.ggobi.org/
GGobi is a data visualization system for viewing high-dimensional data and is the next edition of xgobi. It provides a new interface to many of the features of xgobi, built using Gtk, the GIMP toolkit.

OpenDX !!!

http://opendx.org/

VisAD

VisAD is a Java component library for interactive and collaborative visualization and analysis of numerical data. The name VisAD is an acronym for "Visualization for Algorithm Development".
http://www.ssec.wisc.edu/~billh/visad.html

TANAGRA

produces decision graphs and trees. Includes several classification methods. (Win). Shareware http://eric.univ-lyon2.fr/~ricco/tanagra/en/tanagra.html

LEDA ****

http://www.algorithmic-solutions.com/enleda.htm Library of efficient data types and algorithms (LEDA) contains all of the relevant algorithm building blocks in an easy-to-use and efficient form dealing with objects such as graphs, sequences, dictionaries, trees, points, flows, matchings, segments, and shortest paths.

GDToolkit

http://www.dia.uniroma3.it/~gdt/

AGD

http://www.ads.tuwien.ac.at/AGD/
A LEDA-based library of C++ classes for graph drawing.

Link Lists:

Information Visualization Resources on the Web http://graphics.stanford.edu/courses/cs348c-96-fall/resources.html
(also see InfoVis Cyberstructure above for overview of research/software)

Papers:

Topical clustering, summarization and visualization http://www-nlp.stanford.edu/courses/cs224n/2003/fp/millersj/cs224nfp.pdf
"The growing collection of academic papers available on the web provides a terrific resource for researchers. With this volume comes a need for organization, a task which CiteSeer and other sites have attempted to tackle. We set out to extend the functionality provided by CiteSeer through an investigation of clustering by topic similarity, experimenting with visualization and summary of clustering results. We had three main goals in mind. First, we hoped to develop a sensible clustering of the articles based on textual similarity. In doing so, we ended up using CiteSeer.s similarity results as partial input to our clusterer. Second, we hoped to provide a meaningful visualization of some of the topics represented in the article corpus. Finally, we wanted to investigate the possibility of automatically naming each cluster given the features present in each article and the article.s similarity to the closest centroid. With such a large pool of articles, we felt that automatic generation would be preferable to a hand-crafted taxonomy." May not scale. Some interesting notions.

Mapping topics and topic bursts in PNAS
Ketan K. Mane and Katy B�rner
PNAS, April 6, 2004; 101 (Suppl. 1)
http://www.pnas.org/content/vol101/suppl_1/
Scientific research is highly dynamic. New areas of science continually evolve; others gain or lose importance, merge, or split. Due to the steady increase in the number of scientific publications, it is hard to keep an overview of the structure and dynamic development of one's own field of science, much less all scientific domains. However, knowledge of "hot" topics, emergent research frontiers, or change of focus in certain areas is a critical component of resource allocation decisions in research laboratories, governmental institutions, and corporations. This paper demonstrates the utilization of Kleinberg's burst detection algorithm, co-word occurrence analysis, and graph layout techniques to generate maps that support the identification of major research topics and trends. The approach was applied to analyze and map the complete set of papers published in PNAS in the years 1982-2001. Six domain experts examined and commented on the resulting maps in an attempt to reconstruct the evolution of major research areas covered by PNAS.

Visualization of semantic metadata and ontologies. Paul Mutton and Jennifer Golbeck. In Seventh International Conference on Information Visualization (IV03), pages 300-305. IEEE, July 2003.
http://www.cs.kent.ac.uk/pubs/2003/1655/index.html

Parameters to Visualize Expert Interactions in the Crawl: Help in Lifting and Distillation

Persona: A Contextualized and Personalized Web Search
http://www.hicss.hawaii.edu/HICSS_35/HICSSpapers/PDFdocuments/DTDMI01.pdf

H, Chang, et al
Creating Customized Authority Lists
http://citeseer.ist.psu.edu/chang99creating.html

PEBL: Positive Example Based Learning for Web Page (Make Corrections) Classification
Using SVM Hwanjo Yu Department of Computer Science University...
http://citeseer.ist.psu.edu/564419.html

iVia