Screenshot of FlexPlorer “Columbia.edu” query, try it yourself at http://www.erational.org/software/flexplorer/index.html by entering a URL and displaying and browsing website structures (9)

The idea of using network graphs as navigation tools is not new, and as a note of caution for getting excited about what advantages they could have to offer, information retrieval specialist Marti Hearst writes in her 2009 book Search User Interfaces :
The biggest problem with network graphs is that they do not scale well to large sizes — the nodes become unreadable and the links cross into a jumbled mess. Another potential problem with network graphs is that there is evidence that lay users are not particularly comfortable with nodes-and-links views (Viégas and Donath, 2004), although there is evidence that people can understand networks in which their own social network is the focus of the visualization (Heer and Boyd, 2005). There are several common interactive methods for making large networks more understandable. One idea is to allow the user to select a node of interest and make that node the focus of the display, re-arranging the other nodes around it. This approach can be helpful for medium sized graphs (Yee et al., 2001), but quickly becomes too complex when graphs exceed about 100 nodes. Another approach is to eliminate most nodes other than those surrounding the most recently selected node [1] .

Interface designers who want to use network graphs should certainly keep Hearst’s warnings in mind.//

IBM SmallBlue Network Data Mining and Ex

pertise Discovery Platform


The SmallBlue client is social-sensing software that resides on a registered user’s machine to capture privacy-protected data for social network and expertise inference. It periodically updates new social activities and extracts features from the captured data. The SmallBlue suite features a large-scale social network visualization and analysis tool. For a given topic search, it shows the links among experts and identifies key influencers and brokers. For a given group search, it visualizes the set of people who have common interests. In addition, it can cluster people to find how they interact, and can show the informal group structure.
Figure 30. IBM SmallBlue presentation: Screenshot of Ego network, in which people are sorted by topic area and social distance, which is determined by things like e-mail traffic

Users can search experts within a business division, country, community, group, or specific “social distance,” defined by algorithms that take into account communication activities and collaboration. In addition, there is an interface that allows users set several predefined ranges in tabbed pages that facilitate viewing top experts among various search ranges through only one search submission. The goal of SmallBlue network analysis is to identify influential and informative “nodes” and rank them by their influence in the network and the novelty of information they contribute.
Figure 31. IBM SmallBlue: Screenshot of Expertise finding platform through which experts are suggested
Figure 32. IBM SmallBlue presentation

News Dots (Slate)

How News Dots works
Step 1: Behind the scenes, News Dots scans all articles from major publications—about 500 stories a day—and submits them to Calais, a service from Thomson Reuters that automatically "tags" content with all the important keywords: people, places, companies, topics, and so forth. Slate's tool registers any tag that appears at least twice in a story.

screenshot of News Dot visualization, scan of news stories semantically tagged by Reuters' Open Calais Service

Step 2: Each time two tags appear in the same story, this tool tallies a connection between them. For example, a story about a planned troop increase in Afghanistan reform might return tags for President Obama, the White House, and Afghanistan. These topics are now connected:
external image newsdotsexample.png
Step 3: As this tool scans hundreds of stories, this network grows rapidly, and "communities" begin to form among the tags. Subjects that are highly connected—those that appear together in many stories—cluster together in the network. This occurs in the same way that a picture of the social network of your Facebook friends would reveal clusters of friends from high school, college, and work, with some unexpected connections between them when friends belong to multiple cliques.
Step 4: The news network that results is visualized using Slate's custom News Dots tool, which is built using an open-source Actionscript library calledFlare. Tags are displayed if they appear in at least four stories, and connections are made if at least two stories link those two subjects. The visualization covers the previous three days of news and is updated daily.

BBC SearchMap

"Ice Age" Visualization

Union of International Associations (UIA) Hypergraph search result viewer

UIA Hypergraph view of query for "air pollution" in "world problems" category


Media Giants Links between boards of directors


Orgnet screenshot: executives interacting with advisors


Instructions for the below interactive embedded view of Mr. Bales' Sciologer Pubmed Network Visualization:

  • make sure the Google Earth Plugin is installed
  • use zoom, pan and tilt tools to view the network
  • click on entities for label

Source: Screenshot of Michael Bales' Network in Sciologer

As part of his 2009 doctoral thesis at Columbia University’s Bioinformatics Department, Michael Bales developed Sciologer, a tool for visualizing and understanding scientific communities.

  • The features of Sciologer which, taken collectively, distinguish the system from others, are:
  • Places emphasis on visualization to enhance perception, promote comprehension, and trigger insight
  • Supports querying as data input modality; querying is a familiar action for novice users
  • Downloads records from a database; does not require input of tabular data
  • Uses fully automated, generative network approach; does not require specialized knowledge to create networks
  • Transforms tabular data to the relational data format required for social network analysis
  • Employs a linking schema to specify rules for linking nodes
  • Displays nodes of multiple types (e.g., author, paper, journal, etc. – multipartite networks)
  • Uses RGB cuboid to assign colors based on node position
  • KML output format supports interactive features including zoom, pan, and linkout (hyperlinks)
  • Generalizable to any data set that includes one or more attributes (fields) with categorical data

Sciologer Map Legend

UOC Networked Knowledge — Universidad Oberta de Catalunya

Figure 34. Bestiario: UOC Networked Knowledge Semantic Network Browser based on tags


UOC Networked Knowledge is a forum for interaction with online content linked to the Universidad Oberta de Catalunya (Open University of Catalunya). It allows you access to the University's main resources for teaching, research and knowledge dissemination. To access these resources, UOC Networked Knowledge offers semantic-based browsing, linking the meaning of different contents. This makes access to the intuitive, based on their contextual similarity or proximity, and not from a logical or linear perspective. The basis to this knowledge forum are the tags associated to each information resource.

Visual Understanding Environment (VUE) -

Tufts University

The Visual Understanding Environement, is a free and open source tool for visualizing data. This video tutorial looks at how to use VUE to analyze the contents of a web page and begin constructing a map based on that analysis, bringing in various resources including your Zotero collections, RSS Feeds, and data from other repositories like the New York Times Article API, ad Flickr.

Phylo XML

This displays a Bcl-2 gene family tree annotated with gene duplications, support values, and taxonomy data

phyloXML [example] is an XML language designed to describe phylogenetic trees (or networks) and associated data. PhyloXML provides elements for commonly used features, such as taxonomic information, gene names and identifiers, branch lengths, support values, and gene duplication and speciation events. Using these standardized elements allows interoperability between various applications and databases. Furthermore, both due to extensible nature of XML itself and the provision of <property> elements by phyloXML, extensibility as well as domain specific applications are ensured. The structure of phyloXML is described by XML Schema Definition (XSD) language.


  1. ^

    M. Hearst,
    Search User Interfaces//, Cambridge ; New York: Cambridge University Press, 2009.