Spring Research Conference at BYU

Here are the slides form the presentation that I gave at the Spring Research Conference at BYU last weekend.

Network Roles

Social networks tend to have people that fill positions within the community. For example, within an academic community the roles of being a professor or a student can sometimes be identified solely by using the directed interactions among the individuals. To perform such an analysis, some measure of equivalence is used. In Social Network Analysis by Wasserman and Fauts, the following definitions of equivalence are reviewed (each with a note):

1. Structural Equivalence - requires identical ties to other identical actors
   
2. Automorphic and Isomorphic Equivalence - requires identical ties to other actors
   
3. Regular Equivalence - actors have identical ties to and from equivalent actors
   
4. Local Role Equivalence - actors are role equivalent if they have the same role sets
   
5. Ego Algebra Equivalence - based on algebra of relational structuresSo, why is knowing how to use this important? Well, say you would like to better understand the network surrounding your blog by learning which other blogs are similar in their ties as you, then this is how you could do it.

Blog Posts Increasing and Cyclic

During the span of a year, the over 200 blogs aggregated above show very cyclic behavior. It shows very prominently that these bloggers post significantly more posts during the week rather than on the weekends.

Additionally, as time went on, these blogs as a group posted more frequently.

I would guess that the sink during December was caused by the Christmas holiday.

Security Analysis of Reputation Systems

I came across this report on reputation-based systems today which I found at a reputation based social capital blog. It highlights the security threats against current reputation systems, use cases, and even ten recommendations to combat these threats.

Snapshot of the some of the recommendations

Information Pathways in Social Networks

The first talk presented in the social network session of KDD 2008 was for an interesting paper by G. Kossinets, J. Kleinberg, and D. Watts titled The Structure of Information Pathways in a Social Communication Network (PDF). Although I was not at KDD I was able to watch it online at videolectures.net.
Kleinberg, the presenter, made some interesting observations having to do with our "rhythmic" everyday conversations. The approach to analyzing communication within these social networks is focused on the frequency of correspondence, rather than the content conveyed.

They measure "distance" between individuals by measuring the minimum time required for information to pass from one node to another. A methodology based on Lamport's work and vector clocks in the area of distributed computing.

Using this metric they are able to filter a busy network (one having edges for all communication packets) in a simplified network that contains only the edges that are minimum-delay paths between a pair of nodes. They call this simplified network view the network backbone. Below is an example of such a network (along with the caption) taken from the paper.
The nodes further outside of the center of the graph are more "out-of-date" with respect to node v, since they communicate less frequently.

I found the approach to be novel and useful. As with nearly any analysis technique, caution should be used in selecting the time-period and group size to be studied. Recency and frequency issues come into play as correspondence is aggregated. However, this pursuit offers another approach for more fully understanding information flow.

Facebook growth rising past MySpace

From my local perspective Facebook has been on the rise --- I've noticed that many of my less computer savvy friends have now joined Facebook. I wondered if this trend was global, so I decided to investigate...

During the past few years MySpace has been the dominant social network, however, Facebook has continued to grow much quicker and is expected to become the leading social network. The first plot below (Figure 1) shows a comparison of searches for the keywords "facebook" and "myspace". Lately, for most of 2008, Facebook has been getting a little more attention in the news (lower portion of Figure 1) and has achieved a significantly higher search volume index.

Figure 1. Search Volume Index Comparison of 'facebook' and 'myspace'
(source: Google Trends)

Figure 2 shows the massive popularity of MySpace which began late in 2004, peaked in the middle of 2006, and has since declined --- possibly in part due to the rise of Facebook.

Figure 2. Search Volume Index of 'facebook.com' and 'myspace.com'
(source: Google Trends)

Finally, Figure 3 shows the number of daily unique visitors to Facebook as being more than that of MySpace as far back as November of 2007. (I'm not sure, but I would guess these figures to be based upon Google search result click-thrus)

Figure 3. Daily Unique Visitors of 'facebook' and 'myspace'
(source: Google Trends)

I find it very interesting to see how quickly social networks grow and evolve. As an aside, I think that Facebook is doing things more efficiently and currently providing a better service.

ForwardTrack

ForwardTrack is an open source tool (now entirely written in php) that allows email campaigns to be tracked and mapped as they are forwarded from person to person. This is definitely useful as it reveals the spread of information and some of the underlying social network.

Duncan Watts Downplays Viral Marketing Hype

A while back I quickly saw Clive Thompson's article entitled Is the Tipping Point Toast?, but didn't have the time to read it all nor investigate it any further --- until today.

Thompson's article pits Malcolm Gladwell's thesis (in The Tipping Point) against the recent research of Duncan Watts (cited below). I thought the article was well-written and adequately presented both sides of the issue. In short, Watt's claims that spending time and money marketing to influential individuals is no better than marketing to the masses.

Through all of this, Watts makes some important points such as (quoted from Thompson's article):

• The problem of popular viral marketing talk is that it is "incredibly vague"; "how an influential actually influences is not explained." "Precision matters when trying to explain highly social epidemics"
  
• "Influentials don't govern person-to-person communication. We all do."
  "Common sense is misleading"
• Thompson writes that Watts found the "rank-and-file citizen [to be] far more likely to start a contagion"

So, today I finally took the time to learn more about Watts' recent research, available at Collective Dynamics Group website (at Columbia University) as a Working paper in the Papers section. Through the years, I had previously read some of Watts' work, so I was excited to see his recent findings. In this paper he presents an approach they call "Big Seed Marketing", which in essence combines a traditional mass marketing model with a viral propagation.

The idea that there is "no free lunch" in viral marketing is useful to point out, as "there are many more unsuccessful attempts that one never hears about." He also, points out that it is "hard, if not impossible" to predict which of attempts will succeed.

The take-home message in the conclusion is that effective marketing campaigns can be produced without identifying "influentials", but simply by adding a mechanism of peer-to-peer sharing to propagate the message. (As an aside, the formalism presented in the paper is useful for discussing the problem and easily evaluating the results.)

Watts makes some good points, however, I would still argue that people with high social capital (you might call "highly influential") can heighten the network effect. This is even evidenced in Duncan's paper --- as one of Tom Mauser's 'friend' was StopTheNRA, who, in turn sent a large email blast (Table 1, footnote 1). So, Tom Mauser, had a significant enough relationship with StopTheNRA that they used their resources (their large email list) to forward his message.

Although, there is an element of hype in the presentation of "Big Seed Marketing", I find it useful as it presents a nice way of making the issue sticky and bringing to light these more subtle points. The desired effect of propagating these ideas seems to be occurring.

Update (4/23): Podcast with Duncan Watts on Buzz Marketing (mp3)

Looking for a Job?

There are a lots of places to search for jobs online these days including:

• SimplyHired - easy to use job search
• HotJobs - Yahoo's job site
• Monster - you've probably heard of Monster
• Indeed - newer site that allows you to filter by salary (for instance, here is an example salary search on computer science and data mining)
• LinkedIn - use your social network to find your next job
• erecruiting - for new college graduates

An interesting approach to finding your next job might be to leverage your social connections to match you with a good employer with needs inline with your skills. Of course, as nice that sounds in theory, I would bet it could be challenging in practice.

Although, I won't be needing a full-time job for another couple years, it is always interesting to see what jobs are available (and what skills are in demand) by quickly searching on your skills and interests.

SIP Recap - Thursday

Here is a recap from the Social Information Processing Symposium:

1. Brian Skyrms (UCI), Signaling Games: Some Dynamics of Evolution and Learning
2. John Nicholson (USU), The Blind Leading the Blind: Toward Collaborative Online Route Information
3. Cosma Shalizi (CMU), Social Media as Windows on the Social Life of the Mind
4. Gustavo Glusman (Systems Biologist), Users, photos, groups, words: Analyzing mixed networks on Flickr
   
5. Luc Steels (Vrije U), Social tagging in community memories
6. Aram Galstyan (USC/ISI), Influence Propagation in Modular Networks
7. Adam Anthony (UMBC), Generative Models for Clustering: The Next Generation
8. Peter Pirolli (PARC), A Probabilistic Model of Semantics in Social Information Foraging
9. Hak-Lae Kim (DERI), int.ere.st: Building a Tag Sharing Service with the SCOT Ontology
10. Yu Zhang (Zhejiang U), Mining Target Marketing Groups From Users' Web of Trust on Epinions
11. Andrei Broder (Yahoo), Reviewing the Reviewers: Characerizing Biases and Competencies using Socially Meaningful Attributes (see Sihem Amer-Yahia)
    

The Wednesday talks were excellent. In particular, I really enjoyed:

• The subtleties of the blind leading the blind (see 2 above)
  
• Gustavo's unique way of analyzing Flickr relationships (see 4)
• Adam Anthony's overview of generative models that can be used in clustering (see 7)
• Pirolli's analysis of Lostpedia using LDA (see 8)
  
• Hak-Lae Kim's tag aggregator application (see 9)
• The use of socially meaningful attributes as presented by Yahoo's Andrei Broder (see 11)
  

SIP Recap - Wednesday

I'm here in Palo Alto, California attending the AAAI Spring Symposium at Stanford. So far, the Social Information Processing Symposium has been very interesting and exciting. So far, I've met some people doing some neat research. Today's presentations were the following (I've added links to those I could find online):

1. Bernardo Huberman (HP Labs), Social Dynamics in the Age of the Web
2. Ed Chi (PARC), Augmented Social Cognition
3. Tad Hogg (HP Labs), Solving the organizational free riding problem with social networks
4. Riley Crane (ETH), Viral, Quality, and Junk Videos on YouTube: Separating Content From Noise in an Information-Rich Environment
5. Yi-Ching Huang (NTU), You Are What You Tag
6. Julia Stoyanovich (Columbia), Leveraging Tagging to Model User Interests in del.icio.us
7. Steve Whittaker (Sheffield), Temporal Tagging: Implicit Behaviour Identifies Points of Interest in Complex Event
8. Georg Groh, Implicit Social Network Construction and Expert User Determination in Web Portals
   
9. Elizeu Santos-Neto, Content Reuse and Interest Sharing in Tagging Communities
   
10. Matt Smith (BYU), Social Capital in the Blogosphere: A Case Study (this was our presentation, of course)
    

I enjoyed all of the presentations, in particular I liked Bernardo's address which covered a variety of interesting topics, Ed's comments, Riley's trend analysis, Julia's talk analyzing del.icio.us hotlist generation and tag, Steve's flamboyant presentation, Georg's work (as it had some thoughts related to our work on Implicit Affinity Networks).

(Oh, and I lost my cell phone today.)

I'm looking forward to another great day tomorrow!

Building Community around a Blog

At the spring research conference I presented the following:


















Building Community around a Blog (PDF)