Collab-Or-Ate

Wordle was created by Jonathan Feinberg, another member of IBM Research’s Collaborative User Experience group.

With the election season is in full swing, I’m reminded of this interesting experiment run by Bailenson, et al, at Stanford just before the 2004 presidential election:

One week before the 2004 presidential election, participants completed a survey of their attitudes concerning George Bush and John Kerry while viewing photographs of both candidates side by side (See Figure 1). For a random one-third of the subjects, their own faces were morphed with Kerry while unfamiliar faces were morphed with Bush. For a different one-third, their own faces were morphed with Bush while unfamiliar faces were morphed with Kerry. The remaining one-third of the sample viewed un-morphed pictures of the candidates.

Post-experiment interviews demonstrated that not a single person detected that his or her image had been morphed into the photograph of the candidate. Participants were more likely to vote for the candidate morphed with their own face than the candidate morphed with an unfamiliar face. The effects of facial identity capture on candidate support were concentrated among weak partisans and independents; for ‘card carrying’ members of the Democratic and Republican parties, the manipulation made little difference. [more]

We have more affinity for people we perceive to be more like us and subtle changes to a person’s picture have the power to make us like someone more or less. So be a critical consumer of not just the words but also the images of the candidates! Resist the temptation to vote based on gut feelings about affinity and similarity, because these factors can be easily manipulated.

While at CHI last month (our international conference on human-computer interaction (HCI)), I went to two panels (”Celebrating ‘The Psychology of Human-Computer Interaction’” and “Usability Evaluation Considered Harmful?“) that had really interesting discussions about what defines our research conference (CHI) and our field of study (HCI). I’m still synthesizing my thoughts around these panels and what I’ve been reading since, but based on them, here is how I think about the HCI field today:

1. The HCI field (and the CHI conference) began in the 1980’s with a strong grounding in computer science and cognitive science. Card, Newell, and Moran’s The Psychology of Human-Computer Interaction is considered the seminal textbook describing this approach.
2. Computer science and cognitive science guide us towards taking a systematic, scientific approach to building and evaluating software (for e.g. GOMS). This is a solid way to build systems and many of the early successful HCI research projects utilized this approach.
3. As the software and consumer electronics industries exploded over the last two decades, it has become obvious that there is something more going on here driving user adoption, in addition to computer science innovation and cognitive science usability. You could summarize this as “design” or “context” or “the third paradigm.” However you describe it, it has to do with human emotions, social dynamics and desire.
4. As Greenberg pointed out in his paper presentation, evaluating an early prototype in a systematic way, particularly in terms of usability, can kill the innovation process. Early design often gets things wrong, but it is a critical stage in the product innovation cycle and should not be stunted through rigorous evaluation. He claims that inappropriate evaluation is harming the quality of the work presented at CHI — read Greenberg and Buxton’s paper for more details.
5. The CHI community is struggling to find an identity that simultaneously supports a scientific process (so that there is a criteria for judging quality) and product innovation (so that CHI has an influence over the technology world, outside of academics).
6. The paper The Three Paradigms of HCI (Harrison, S. Tatar, D. and Sengers, P.) tries to define exactly what this “third” thing is that is missing from our traditional HCI education, calling it the “phenomenological matrix.” Research practices this third paradigm include are ethnography, action research, practice-based research, and interaction analysis, where the “goal is to grapple with the full complexity around the system.”
7. Because I’ve been working within the space of design, social psychology, and “context” for so long, this approach to building technology seems so logical, yet surprisingly hard to justify to CHI paper reviewers. But on the other hand, my response should not be to reject the CHI’s body of work as misguided or uninformed. I think a rejection of stringent evaluation techniques should not lead to a rejection of the innovations that have been born out of this structure.
8. My conclusion from this is that I should read more, spending time becoming more aware of and inspired by the work done before us. I’m all in favor of coming up with alternative evaluation methods or no evaluation criteria so that we can foster risky, exciting ideas within HCI. But I don’t want to abandon all the early work’s ideas.

Some Recommended Readings:


The Psychology of Human-Computer Interaction, Stuart K. Card, Thomas P. Moran, Allen Newell

Twenty-five years ago, Card, Moran and Newell’s book, “The Psychology of Human-Computer Interaction”, named our field and launched us into a new world of user-centered design and development. These pioneers believed that “a scientific psychology should help us in arranging [the human-computer] interface so it is easy, efficient, error-free – even enjoyable.”

Saul Greenberg & Bill Buxton’s paper “Usability Evaluation Considered Harmful (Some of the Time).“

Current practice in Human Computer Interaction as encouraged by educational institutes, academic review processes, and institutions with usability groups advocate usability evaluation as a critical part of every design process. This is for good reason: usability evaluation has a significant role to play when conditions warrant it. Yet evaluation can be ineffective and even harmful if naively done ‘by rule’ rather than ‘by thought’. If done during early stage design, it can mute creative ideas that do not conform to current interface norms. If done to test radical innovations, the many interface issues that would likely arise from an immature technology can quash what could have been an inspired vision. If done to validate an academic prototype, it may incorrectly suggest a design’s scientific worthiness rather than offer a meaningful critique of how it would be adopted and used in everyday practice. If done without regard to how cultures adopt technology over time, then today’s reluctant reactions by users will forestall tomorrow’s eager acceptance. The choice of evaluation methodology - if any - must arise from and be appropriate for the actual problem or research question under consideration.

The Three Paradigms of HCI, S Harrison, D Tatar, P Sengers

Informal histories of HCI commonly document two major intellectual waves that have formed the field: the first orienting from engineering/human factors with its focus on optimizing man-machine fit, and the second stemming from cognitive science, with an increased emphasis on theory and on what is happening not only in the computer but, simultaneously, in the human mind. In this paper, we document underlying forces that constitute a third wave in HCI and suggest systemic consequences for the CHI community. We provisionally name this the ‘phenomenological matrix’. In the course of creating technologies such as ubiquitous computing, visualization, affective and educational technology, a variety of approaches are addressing issues that are bad fits to prior paradigms, ranging from embodiment to situated meaning to values and social issues. We demonstrate the underlying unity of these approaches, and document how they suggest the centrality of currently marginal criteria for design, evaluation, appreciation, and developmental methodology in CHI work.

Thomas Erickson, David W. McDonald’s new book, HCI Remixed: Reflections on Works That Have Influenced the HCI Community

From Tom Erickson’s web page:

The goal of the HCI Remixed project is to produce a collection of essays in which researchers and practitioners reflect on a paper or other piece of work by someone else, that is at least 10 years old, and that has had a personal impact on their view of or approach to HCI.

It is that type of day where you mean to get some work done, but you first have to check out your blogs. Because checking blogs is kind of like getting work done, right? So of course I’ve spent half of Saturday catching up on politics and pop culture and not getting any work done. Unintentionally though I think I found something to help with my data analysis, which I swear I will get down to as soon as I finish this blog post!

On Lynn Cherny’s blog post about Colbert and book sales, I was admiring the data charts she posted from Juice Analytics. Turns out they have a bunch of tools and tips for making your chart-junky Excel charts look a wee bit better with a simple click of the button. I’ve downloaded their CleanCharts Excel Add-on and, when it works (kind of quirky), it works great. I also picked out the best of the charts in their Chart Chooser and downloaded those templates.

For my own research purposes, I’m using my fellow researchers’ tool, Many Eyes. Pretty cool!

And I need a place to save this, so why not use the blog?

I was also presenting a poster at Group on a project I worked on while at Sun Microsystems Laboratories. The application is called Constellation and it uses social network visualization techniques to reveal to users the location and relationships between experts and novices in an organization.

Joan DiMicco, Nicole Yankelovich. (2007) “Constellation: Using Visualization to Find the Path to Experts.” Poster Presentation at the ACM Conference on Organizational Computing and Goupware Technologies (GROUP 2007), Nov 2007.

This image is a screenshot of Constellation. The nodes are people and the lines are the relationships. The social network relationships shown are the management structure, co-authoring history of internal and external documents, patenting history, and neighboring offices. The blue lines (the thicker lines) represent multi-dimensional relationships, meaning the pair are connected in multiple ways, such as both authoring and patenting together. The user can turn any of these relationships on/off at will.

What I think is cool about Constellation is that you can figure out the relationships between experts in a topic area, for example here the experts in “hardware” are shown in the screenshot. There are two hardware research clusters in Sun Labs that have done a lot of work together, but haven’t done any collaboration across each other. The weak linkages between these two clusters of researchers is a tie that says “sits near.” So if you wanted to begin to have these groups work together (by writing papers or patents together), the first step is to go to the people who have the “sits near” relationship and get them to start talking to each other.

The proposed use of Constellation is for novices. If I add myself (a novice in “hardware”) to the graph, I can see the shortest relationship path between myself and the experts already shown. By exploring the relationships between these experts and myself, I can figure out the best way to get a personal introduction to an expert in hardware.

Users have so much knowledge about who they know and who they comfortable asking for help from that is NOT captured electronically, that I think the strength of Constellation is leaving the ultimate decision of who to contact entirely up to the user. Rather than presenting the user with a ranked list of mysteriously calculated “hardware experts,” this interface lets the user visually explore the social network space and selectively focus in on the persons of interest.

(I also wrote about an extension of this project for the CHI 2007 Shared Encounters Workshop, “Enriching Encounters with Social Networks.“)

Here’s a mash-up of Google Maps with the ubiquitous Yet Another Social Networking Service (YASNS), Facebook.com:

MapYourBuddies!

Where did your friends grow up?

(Found on Google Maps Mania.)

Clusty generates a tag cloud for you from any search term. It seems pretty interesting. I found this through Ghostweather.

It is interesting to see what it comes up with from a search on my name — most prominent are my graduate school advisors who have been my most frequent co-authors. I’m impressed it got my current and past employers, but I wish it would drop the proper names entirely and show more descriptive keywords. But anyway, kind of a fun little widget to put out there.

It is Friday afternoon and I can’t help but think, where has the week gone?? How many times have you asked yourself that?

If you are on a Mac, there’s a new application you can use to track what you do on your machine throughout the day. Maybe it can help you get a better answer to the eternal question “where does the time go?” It can definitely show you some interesting visualizations of how and when you use different applications on your machine.

The tool is called Slife, from Slife Labs. From the website:

Slife is a new application for the Max OS X that lets you visualize and organize your computer activities like never before. Slife observes your every interaction with applications such as Safari, Mail and iChat and keeps tracks of all web pages you visit, emails you read, documents you write and much more.

From what I keep reading and seeing from the “MySpace” generation, there is a strong desire to share status with friends, family, and even loose contacts. And Slife Labs has an add-on tool to allow you to share your Slife captured actions and publish them to your contacts. So if you want to share more information with your friends about how you are spending your time, SlifeShare offers you a new way to do that:

A Slifeshare is an online space where you share your digital life activities such as browsing the web and listening to music with your friends, family or anyone you care about.

It is all very interesting stuff and reflects the current trend of capturing and sharing life data.

If you don’t work at IBM, you probably are not aware that this week is Lotusphere.
As an IBM and Lotus Notes newbie, I’m only just beginning to understand the history and culture surrounding the IBM Lotus products. Basically, Lotusphere is like Macworld for Notes users, although admittedly not quite as cool. (Sorry, no iPhone, but we got SameTime 7.5.1!) Because the Lotus software products are focused around collaboration and workplace productivity, these products are the ones most closely related to my research and the research of the other members of the Collaborative User Experience group.

This year’s Lotusphere has generated a lot of press (I don’t think this is usual), but it is fun to see press about the research projects of my colleagues.

InformationWeek has an nice article covering the coolest research that will hopefully become Lotus products:

One of the best things about the IBM Lotusphere conference is always the glimpses it gives you of the future of computing. The various IBM Research labs send representatives who staff a room filled with demo pedestals — two dozen this year — where creators show off their projects. This year, as usual, several projects look like good prospects to become future products, and IBM Lotus has even put one up on the Web so you can get a look at it even though you’re not at the conference.

The article mentions Malibu, Tagging in software development, and Many Eyes, all from CUE.

This is the kind of 3D visualization I 100% endorse! The 3D tangible objects nicely illustrate the scale of the data and they are delicious to eat.

The U.S. Defense Budget, Explained with OREO Cookies

The New York Times published an article today on the difficulties patients have in making appropriate decisions about their medical care: In Medicine, Acceptable Risk Is in the Eye of the Beholder. It is very hard for people to understand the risks associated with medical treatments, because as humans we are particularly averse to risk when it comes to potential losses to our personal situation. What can be more personal than one's health? And as the article discusses, patients frequently make sub-optimal decisions for themselves because of this aversion.

To read more about this “irrational” behavior, start by looking into prospect theory, a seminal idea from the field of behavioral economics, developed by Kahneman & Tversky.

This article stands out for me though because it explains how visualizations of the risks can dramatically help patients understand the issues more completely. I am very interested in this integration of visualization & decision-making and would like to see more systematic analysis of why different visualizations lead to different conclusions.

From the NYTimes:

In a paper published in the June issue of PLoS Medicine, Dr. Jerome R. Hoffman says using illustrations is helpful. Pie charts, dartboards and, best of all, roulette wheels, he suggests, communicate the complex information about the probability of a good outcome more understandably.

My question is why are roulette wheels the best visualizations? Is it because, on average, patients are most familiar with the concept of risk from gambling with roulette wheels? Does that mean you should get a visualization tailored to your personal life experiences? Nerds get pie charts, barflies get dartboards, and gamblers get roulette wheels? Or is there something inherent in our visual analysis that is universal for all types of people, making roulette wheel visualizations easier to analyze than the other representations?

For inspiration on how to present lots of data in a playful and informative way, check out this applet We Feel Fine. (found on information aesthetics, as usual.) It visualizes blog entries that refer to feelings, organized by time. The content is interesting enough, but I most enjoyed the non-traditional ways of presenting statistics and large amounts of data.

Here are some screenshots to get you interested in checking it out:

Selecting a random point in a large sea of data:

Ordered list of term frequencies:

Histogram with many datapoints missing:

Histogram with organic, animated blobs:

Due to the insights of our commentors, I've changed my mind about circular treemaps. They have a distinct disadvantage compared to rectangular treemaps in that the parts (sub-circles) do not add up to the whole (the surrounding circle). But circular, colorful data representations still appeal to me.

Ciros circular graphs (shown above) are quite appealing. They represent a method for visualizing genome relations, but could be applied to any complex relational dataset. It is also an free GPL project, so you can try it out for yourself. For inspiration, checkout these lovely screenshots.

For those of you who are big fans of circles (me! me!), there's a new treemap visualization on the street: “Circular” Treemaps. They look more organic than Ben Shneiderman's original Treemap visualizations.

Circles:

v. Squares:

found via information aesthetics

— Next Page »