Teaching

On this page you will find teaching materials developed for modelling and analysis of social networks. My teaching curriculum includes taught masters, masters-by-research and conference  tutorials. As these are usually practical sessions taught in the computer lab, they were easy also to convert into web-tutorials.

Workshop on Agent Based Modelling and Social Network Analysis

Richard Taylor and Gindo Tampubolon

Offered by the ESRC Research Methods Festival in association with the Oxford Spring School held by the Department of Politics and International Relations, 19th July 2006

Course Details

Both agent based modelling and social network analysis are valuable for investigating dynamic processes of social phenomena of interest to social scientists. For instance works on technological  or social innovations and social movements have fruitfully applied these methods. The hands-on workshop will use free software (Java, Repast and Pajek) downloadable from the web.

Participants are encouraged to bring their  problems, data or questions along to be discussed during the workshop. Please bring problems or data in their original setting or format. An important issue to be discussed is how these modeling and network perspectives can be used  to understand the problems and to transform data from its original setting into a format amenable to analysis.

GT's online materials

RT's online materials



Network Analysis with UCINET for ABM of Industrial Districts and Economic Geography

by Richard Taylor

University of Napoli "L'Orientale", February 23-24 2005

Description

This course will look at how to characterise social networks (their structure and their evolution) in 'industry' models of firms' and institutions' interactions. It is a practical, laboratory-based class which  will introduce UCINET software, used in conjunction with other programs and platforms. The tutorial will combine overhead presentations with short practical exercises and supervision. It  doesn't assume any previous knowledge of programming, nor with the specific software used. Principles of Agent-Based Modelling (ABM) will be described very briefly. Social network theory  (the concepts of path length, cohesion, clustering, cliquishness, centrality etc.) is covered in earlier classes, so in this tutorial there will be a short revision. The practical course will be split into two  parts. In the first part, models written in Java and making use of the RePast packages will be used to generate networks data (first day). In the second part the data will be analysed using UCINET (second day).
Duration: 6 hours (2+4).
Related concepts: supply networks, supply chains, Industrial Districts (IDs), geographical simulation, urban simulation.

Course Outline (4 page- doc file)

Slides day1  day2 (pdf)

 Additional material at Analytic Technologies: handouts


RePast

 The Industrial District (ID) network model consists of two types of agent: client and subcontractor firms, and shows the (directed) relations among them. It is initially constructed by linking random  client - subcontractor pairs according to some parameter settings, which can entered by the user via the RePast toolbar. When the "Run" button is pressed, the initial network is then displayed  graphically. Subsequent dynamical steps can then be added: nodes can be removed, and further contracts can be added, whereby the network evolves towards a more complex structure. The  network data (in the form of adjacency matrices) is written to .dl files at the beginning and end of each simulation run.

The s ource code for the tutorial is in two files, IDNetworkModel.java and Firm.java. You will also need the RePast toolkit (the model was developed with version 2.2), Java, and I also recommend UCINET.


Back | Back to the  CPM

 Page last updated 21/1/2007