Someone pointed out this clustering application for del.icio.us. Its just for clustering your own data though, so its not leveraging the group mind. Still it was interesting to see the clusters formed by my tags. Some of the clusters were: clusteranalysis(!); visualization; mobile research; market research; flash; ajax; cardsorting; design.
Expectedly, they are using k-means clustering algorithm. I played around with the number of clusters and the cohesiveness of the clusters does change as the number of clusters goes up.
It was bound to happen sooner or later. Simple lists are great, weighted lists (or tag clouds) are even better. But as data accumulates, its no longer practical to navigate it through tag lists. Some sort of structure is needed. Back in February I had written about tag-sorting, and how easy it should it should be to cluster tags.
Enter categories or clusters as they are referred to on Flickr.
How it works: When you click on a tag, you have the option of exploring different clusters related to the tag. For example, summer has several associated clusters:
Word is Yahoo will soon be coming out with its OddPost inspired version. Ethan Diamond, one of the Oddpost founders, now working on Yahoo mail thinks that users are more likely to use folders than tags to organize emails. I tend to agree. I like the controlled messiness of del.icio.us tags for my bookmarked urls. And the social discovery of others’ bookmarks is very compelling. But for something as personal and important as my email archive – I prefer either a combination of folders and robust search.
The search field is heating up again. There have been a spate of innovations and acquisitions by various search companies. Particularly exciting have been innovations in Local Search, Desktop Search, Maps, and Image finding.
I will be moderating a panel at BayCHI on this very topic on April 12th. Panelists include John Battell of SearchBlog, Udi Manber CEO of A9, and Rahul Lahiri VP, Product Management for Ask Jeeves.
I am talking to two more panelists. Look forward to an announcement soon.
All morning I have been wishing I was at etech. The next best thing is to troll the wiki and see who is posting notes. I even ventured to IRC (#etech) where people are ooh-aahing over Joel’s talk.
Just saw Cory Doctorow’s notes from the folksnomy panel.
This comment by Stewart Butterfield caught my eye:
“The idea that it’s possible to cleave nature at the joints is controversial. Yes, there are countries, Uzbekistan is a country, but ask a physicist or a biologist and the categories are very fraught.”
So marketers have already started advising companies to pay attention to tags. So I started thinking – what would an information architect do with the wealth of information given by del.icio.us / flickr / technorati tags?
The first thing that comes to mind is to use tags as a proxy for free-listing. Information architects or anyone else researching a domain IAs can perform a card-sort on tags instead of generating items for a card-sort exerise using freelisting.
Just noticed a tutorial on card sorting at CHI.. Here is the description:
“This tutorial will provide researchers with practical guidance and hands-on experience in card sorting and cluster analysis methods and tools…
-What is a card sort and cluster analysis?
-How do I perform a card sort and cluster analysis (tools and methods)?
-What else do I need to know to be able to perform this method in practice?”
My question is – why on earth are they restricting themselves to cluster analysis. Although they do not say so explicitly, I suspect they are probably going to mostly talk about “Hierarchical Cluster Analysis”. This is such a restrictive use of Card-sorting which is capable of being a general purpose user mental modeling technique.
We have used Hierarchical Cluster Analysis, K-means cluster analysis and Multidimensional scaling with great success. Different types of problems require different analysis techniques. Architectural problems might require Hierarchical Cluster Analysis, but I often find such analysis too restrictive and prescriptive (it gives me a structure – take it or leave it). In contrast, I find the results of MultiDimensional Scaling helps more in understanding the domain for an exploratory project.
I really need to put together some examples of visual deliverables for different types of card sort analysis. If only my clients would let me get out of those NDA’s, and let me share some examples!
November 4, 2002 — Information architects from across the world today announced the launch of the Asilomar Institute for Information Architecture (AIfIA). The leadership of AIfIA includes expert practitioners, teachers, and authors from organizations including Yahoo!, AOL/TimeWarner, IBM, Lucent Technologies, MasterCard International, Wells Fargo, Wachovia Corporation, Razorfish, Adaptive Path, the Transportation Security Administration, the University of North Carolina and the University of Texas.
AIfIA is a non-profit volunteer organization that serves as a resource for
organizations and individuals seeking to learn more about information architecture and its benefits, and assists information architects who wish to promote the field. Information architecture, the art and science of structuring and classifying information on web sites and intranets, is a growing field that is becoming increasingly important in the modern information age.
“Information architecture leads to increased revenue, decreased development costs, more effective communication and successful web sites,” said Christina Wodtke, AIfIA President and author of the book Information Architecture: Blueprints for the Web. “AIfIA will be a resource for those wishing to learn more about information architecture, whether they’re exploring its benefits for the first time or they’re experienced professionals sharing tricks of the trade with their peers.”
I am happy to tell you that yours truly is involved with this. Take a look at the AIfIA site, read the mission, find out who are the people involved and perhaps you will see why I chose to get involved. Consider joining, volunteering. Also, please tell us what you think of the AIfIA, its mission and plans. We welcome your questions and feedback and will pay close attention to what you have to say while planning the next steps.
For the first time ever (as far as I know) the Nobel Prize in Economics has been awarded to a Cognitive Psychologist. The work of Daniel Kahneman on how people make decisions has had a major impact on both Economics and Psychology. A graduate course I took on decision-making mostly focused on his work and its critiques. Through a series of innovative experiments, Kahneman & Tversky investigated how people deviate from “rational decision making” using heuristics (short cuts) to make their decisions: (a) people pay attention to some relevant information and ignore other information; (b) are affected by framing (how the information was presented); (c)judge relationships between presented information incorrectly; and (d)ignore base-rates (probability of event happening). (a good book for non psychologists to explore this area is “The Psychology of Judgment & Decision Making“)
Why is this relevant to the design of interactive systems? Many interfaces we create support human decision-making. And a good system should take into account how human beings make decisions. For example, Collaborative Filtering Systems (e.g., Amazon type recommendation system) are typically suited to the idiosyncrasies of human decision-making (though they have nothing to do with Kahneman’s insights). People are used to relying word of mouth, to social filtering of information. And Collaborative Filtering algorithms use this same characteristic to create a new paradigm for finding information online.
I believe that as interfaces continue to evolve one of the major developments will be incorporating an understanding of how people make decisions. This will happen as researchers in Universities launch programs for understanding how bounded rationality helps explain online decision making. Academic research will mostly involve lab-based experiments (in Business Schools, Behavioral Economics and Cognitive Psychology Departments), but will also be aided by server log analysis. Results of such research will ultimately permeate the HCI community and influence the design of ecommerce systems, search interfaces and online auctions etc. (Every once in a while I feel tempted to do such research myself!)
In the meantime here are some links about current research on online decision-making found after some quick googling:
(a) Conference on Online Consumer Psychology
(b) A paper on Consumer Decision Making in Online Environments (pdf)
(c) BOunded Rationality and Satisficing in web-based decisions (JASIST paper)
(d) Emotions in Decision Making (recent research has explored this role)
(Note: CogPrints and is a good sources for online articles on Psychology and Cognitive Science.)
Design personas are one of the most powerful tools in a designer