Small World Networks 2008
Contents
Supervisors
Prof Derek Abbott & Dr Matthew Berryman
Weekly progress
Week 2
Wenrui (Linda):
Zhujing (Sweet):
Project proposal
For the project proposal you need to have:
- Some fun stuff (fun small-world networks)
- Some more serious engineering reasons why small-world networks are important - namely robustness of power networks, communications networks, Google searching, finding who is the leader in social networks (like we discussed about Enron)
- Some slides on network measures (correlation coefficient, cluster detection algorithms).
- How you will start with some small networks that you can test your software on before moving on to a large data set.
- Some slides on project management - how you might structure your software modules & data flow, and definitely Gantt Chart (including slack space to cope with risks), budget, milestones, who is doing what (& why).
Here are some fun network graphs: [1] There are some fun and some more serious networks here: [2]
When you include graphics from the web, don't forget to include on the slide the web address you found them on. It is also a good idea to have a slide at the end with some key references (e.g. Barabasi book, Milgram's paper, the paper I gave you on finding hierarchies).
Critical Design Review
For the Critical Design Review (not due until Week 8) you need to have the following parts:
- Project Aim & Background
- Literature review, where you discuss a number (say 10-15) of papers / books and their contributions. Try and summarise in a few sentences what each one is about.
- Your approach - what you are going to do, your software design (see below), and who is going to do what.
- Analysis of your software design - what are the hard sections, what are the critical parts that you need to get working first.
- Project management stuff. For risks you don't need to worry much about Occupational Health & Safety (apart from general computer set up) but there are other risks in your project to do with task timing, what happens if someone (including Derek and I) is unavailable, budget over-runs (should be minimal risk but you can still mention it).
Aside from the engineering applications (where small-world networks are important), the other engineering in this project is software engineering. You should have a block diagram showing the main parts of the software, i.e. graph file format converter, graph drawer, network measure calculators, hierarchy algorithm, and matlab analysis (graphing and statistics), and my software for converting the email data set into a graph file. For some of these you can then include what sub-modules they have, e.g. the network measure calculator might have sub-modules for degree distribution, for clustering coefficient. You should also have a flow diagram showing the data flow between these modules, and in the critical design review give details of the algorithms. You can software you find on the web but you will need to acknowledge it in your proposal report, make sure you understand how it works, and test it. It may be easier though, to write the software yourself then writing a file format converter to get other software to work.
The Critical Design Review can just be emailed directly to myself and Derek on the due date, and you only need one for your group. For the final report you need to write one each (please don't copy each other's writing but write it in your words - you can copy equations, flow charts, graphs, diagrams, reference list, software though. Derek and I are happy to read draft copies.
Reading list
Data sets
- Enron email data
- Network data compiled by Mark Newman. Most of the data is in the GML format. This data includes the karate club social data.
