SI 508 Networks: Theory and Applications

Resources

SI508 home

cTools site

assignments

software tools

related courses


Schedule

Fall 2007:

Tuesdays
508 & 708 together
4-5:30pm in WH412

Thursdays
708 10:30am -12pm in the Complex Systems Commons (WH 317A)

508 4-5:30pm in the DIAD lab

Office hour:
508/708 Mondays 3-4pm




SI 508 & 708 projects

 

 

This page now serves the purpose of showing off students' projects.

This course is now part of University of Michigan's Open Educational Resources initiative. You are free to use these materials (I encourage you to do so), but please use the OER version.

Click here for the syllabus of the course.
If that's not exciting enough for you, you can experiment with some of the online demos.

World of Warcraft network visualization

Wow! We’re doing WoW
Matt Rubinstein, Yarun Luon, Sameer Halai, and John Suciu

An analysis of over 100,000 users, 5,000 guilds and 15,000 arena teams in World of Warcraft (a massively multiplayer online game), looks at whether position in the network correlates with performance in competitive play.



Influence and Similarity between Contemporary
Jazz Artists, plus Six Degrees of Kind of Blue/
Gabriele Giaquinto, Cora Bledsoe, and Brian McGuirk

Gabriele, Cora and Brian first figure out whether they can map contemporary musicians to jazz subgenres, using a similarity network. Then they trace the current musicians who have assimilated the most direct and indirect influence. Finally, they examine in detail the 1959 jazz collaboration network, with an interesting finding: those playing rhythm section instruments act as connectors in the network.


Exploring Best Practices for Image Digitization
Liz Blankenship, Chen Huang, Alok Vimawala

Prof. Paul Conway had 17 documents and a question: how did image digitization best practices come to be? Through some creative network visualization and analysis, Liz, Chen and Alok figured out who the key individuals and institutions are, and how the documents and people relate to one another over time.

Network Structure, Diffusion and Growth Over
Time in a Facebook Application

Hung Truong

Hung wrote a Facebook app to allow people to take and share notes. He then studied the adoption of the app as it spread, in part, through the social network.


Discovering Salience in Textual Elements using
Graph Mutual Reinforcement

Ahmed Hassan

Ahmed came up with an elegant idea: sentences that provide the best summary of a piece of text have important words. And important words are those that are contained in important sentences. Sounds like a job for the HITS algorithm, and Ahmed takes it for a spin.


Comparing the networks resulting from navigational and information queries
Valentina Karimova and Ahmad Aljadaan

Valentina and Ahmad had a lot of intuition on how navigational and informational queries might differ, but very little has been done to understand how the search results differ for these types of queries. by constructing query connection subgraphs, they find that navigational queries will have more densely connected search results and more pronounced hub structure.


A social network analysis of YouTube
Ping Yu, Min Hu, and Nayeoung Kim

Precisely how social is YouTube? Do users with common interests form densly linked communities? Not so, find Rebecca, Min and Nayeoung. Neither the friend nor subscriber networks are especially dense - leaving one to wonder whether the social interactions occur through comments left on videos instead...

Community Finding in a Lexical Network of News Feeds
Jessica Hullman and Bryan Gibson

Can network-based clustering algorithms outcompete more traditional hierarchical clustering approaches when it comes to grouping news articles according to content. Jessica and Bryan find that, network clustering still has a ways to go.

Network Public Goods
with Probabilistic Exclusion:
Mechanism testing with agent-based modeling
and experimental design

Gregory Gamette

Will excluding non-cooperating players probabilistically in a network game lead to acceptable outcomes for the network as a whole? Greg explores using agent based modeling and econ lab experiments.

task prestige diagram

Competing to Share Expertise: the Taskcn Knowledge
Sharing Community
Jiang Yang

Witkeys are a type of website in China where users can pose tasks (with monetary reward) and others compete to complete those tasks. Jiang applied network analysis techniques to identify experts, and answered questions such as: Does offering more money attract more attempts by better experts? Can one predict the chances of a user winning a task based on their past performance and who is in the competition?

Patterns of US Internal Migration in late 90s
Anya Polukeyeva

Some people moved from A to B while others moved from B to C. What happens when you connect all these flows? Anya uses network analysis to find what the actual regions of the US are among which population circulates, and to find trends along ethnic and socioeconomic boundaries.

What’s good for me is good for us: The effect of
network structure on using social psychology to
motivate contributions to a social network referral tool
Erin Gong

The following projects include interactive models and demos that you can try out yourself

Network Structure Matters
A Network Analysis of Micro-loan
Borrowing Groups
David Hutchful
Pieter Kleymeer

Whether an individual repays a microloan depends on their cultural and geographical proximity to other individuals in their lending group. David and Pieter explore the parameter space with an agent based model.
Download the model (you'll need NetLogo to run it)

 

A network analysis of STI transmission in two different
Tjimba/Himba communities

Ashley Hazel

What will happen to STI rates an isolated tribal community as they come in contact with more and more outsiders? Ashley investigates using agent based models.

Download the Ombivango model (you'll need NetLogo to run it)
Download the Otjikoyo model



Modeling an Agent-Based Decentralized File Sharing
Network

Alex Gonopolskiy and Benjamin Nash

How much freeriding can a peer-to-peer network tolerate? Alex and Ben investigate a filesharing protocol where files are shared piecewise.

Download the model (you'll need NetLogo to run it)

An Extensible Graph Toolkit for Mathematica
Eytan Bakshy

Eytan develops a fast, flexible, and powerful toolkit for network analysis in Mathematica

Download the mathematica file (you'll need Mathematica run it)

more to come (I still need to put up a couple of more that I do have permission for) + others I hope to link to once their authors have submitted them for publication... stay tuned!