Computer Programs for Social Network Analysis

There are new visualization tools around us emerging all the time designed for

research use to map our social networks. Although they are often based around

innovative－and sometimes unfamiliar－methods and measures, it is still well worth

checking them out. We will introduce some popular social network analysis software

and try to be clear about what they are trying to do. The following programs are in no

particular order or priority.

A. UCINET

UCINET was produced by a group of network analysts at the University of

California, Irvine (UCI). The current version is 6.41 and development team is Stephen

Borgatti, Martin Everett and Linton Freeman (2002). It is a comprehensive package for

the analysis of social network data as well as other 1-mode and 2-mode data. It can

read and write a multitude of differently formatted text files, as well as Excel files. It

contains social network analysis methods including centrality measures, subgroup

identification, role analysis, elementary graph theory, and permutation-based statistical

analysis. In addition, the package has strong matrix analysis routines, such as matrix

algebra and multivariate statistics.

Figure 2.10 A screenshot of UCINET for windows

B. KrackPlot

KrackPlot is a program for network visualization designed for social network

analysts. It was devised by David Krackhardt, Jim Blythe and Cathleen McGrath

(1994). The program is distributed by Analytical Technologies. The current version is

3.3. It runs on DOS systems. Data input files can be produced by converting UCINET

files or directly in a text editor. It is specifically designed to draw data on screen and

can produce circle diagrams and MDS (Multidimensional Scaling) displays. Points can

be labeled and they can be moved under mouse control. Sociograms can be displayed,

printed, or saved in GIF format.

Figure 2.11 A screenshot of KrackPlot

C. Pajek

Pajek – the word is Slovenian for spider – is a program for large network analysis.

It was produced by Vladimir Batagelj and Andrej Mrvar at the end of 1996 and has

been periodically updated. An improved version of the paper presented at Sunbelt’97

was published in Connections by Batagelj and Mrvar (1998). As development is still in

progress, it is still not in official release form. The current version is 0.99. The program

is a Windows program that displays its results and workings in a main window and

various subsidiary windows. It allows some powerful processing of large networks and

is easy and accessible to produce network visualizations.

Figure 2.12 A screenshot of Pajek

D. StOCNET

StOCNET is an open software system for the statistical analysis of social

networks using advanced statistical models. It was produced by Boer, Snijders and

Zeggelink in 2000. The current version is 1.4 (Boer, Huisman, Snijders, & Zeggelink,

2003). StOCNET provides a platform to make a number of statistical methods. The

three main goals are: First, to be the incorporation of important recently developed

methods for statistical modeling of social networks in user-friendly and

easily-available software. Second, to be more efficiency in the implementation of new

methods by setting up a system with common data structure and user interface. Third,

try to be faster availability of new methods.

Figure 2.13 A screenshot of StOCNET

E. NetVis

The NetVis module allows a dynamic visualization of social networks. It is a free

open source web-based tool designed to simulate, analyze, and visualize social

networks using data from csv files, online surveys, and geographically dispersed work

teams. It was produced by Jonathon N. Cummings in 2002. The current version is 2.0.

One characteristic is to visualize social networks with 3D MAGE program. It allows

users to download different data formats of social networks such as UCINET and

KrackPlot. It also includes a simple 2D GraphLayout display for the social networks.

Figure 2.14 A screenshot of NetVis module

2.4 Summary

In sum, the aforementioned sections of literature review suggest that visualization

of social networks using social network analysis is a good approach to observe

communication, interaction, and relationship among people in an online community.

First, all terms of social network analysis should be obeyed in the research. Several

attributes, e.g. size, degree, desity, etc., of social network analysis were implemented

to analyze the sociograms. Second, according to the three types of display in web

browsers, the Spring Embedder of Java 2D Applet were chosen to display the

sociogram of social networks in the online discussion forum. The reason is that

participants in the forum can see the sociogram easily, intuitively, and in real time. As

for 3D display of VRML or MAGE, we considered that it is not essential after all. 3D

display might complicate the view of sociogram though it is fantastic and splendid. It’s

too bad that will violate our goal of simplification in visualization of social networks.

Last, there are many programs for social network analysis. The new ones are

always appearing too. But how do we choose these programs to use? The choice of

which of the main programs to use might be a matter of personal preference, and,

perhaps, of personal finance. Through careful considerations, we finally decided to

create our own programs for visualizing social networks though there are some

programs lying there waiting for us. The main reason is that the one we developed can

be just appropriate for our own particular demand such as web-based and easy to use

for participants. Some of the programs or packages such as UCINET, Pajek, and

StOCNET, etc. are powerful and for different purposes, and we think they will be very

helpful in the future work about the more and deeper analysis of large datasets in social

networks.