Designing Visualizations of Social Activity: Six Claims

posted on Jul 17, 01:01 PM in Social Software, Design

Erickson, T. Designing Visualizations of Social Activity: Six Claims, The Proceedings of CHI 2003: Extended Abstracts. In press. New York: ACM Press, 2003.

This is a wonderful paper by Thomas Erickson of IBM’s T.J. Watson Social Computing Group that puts forth a set of six “claims” on how to successfully design Social Activity Visualizations. The six claims are:

  1. Everyone sees the same thing; no customization
  2. Portray actions, not interpretation
  3. Social visualizations should allow deception
  4. Support micro/macro readings
  5. Ambiguity is useful: suggest rather than inform
  6. Use a third-person point of view

Though Erickson discovered these claims through his work on “Babble”, a wonderful, if not widely used, conversational interaction support software, many of them have deeper significance for the general field of Social Visualization.

The first step in understanding why these claims are true is to recognize that the design of a successful Social Visualization and interaction interface must build upon users’ physical world expectations. One of my favorite, if not obvious, statements about social software systems is that they aren’t doing anything fundamentally new—we are, after all, social beings and have been for millenia. You must always question how a person would perform an action or expect to interact in a physical world situation, and then attempt to translate that action or expectation into an online equivalent.

Indeed, if we look at some of these claims from a deeper, human interaction level, their validity becomes apparent. Take for example claim #1. In a simple physical world interaction, person A sees and is seen by person B, and vice versa (interaction between people prevents one of those people from hiding). Additionally, they both see their environment, and except for a slight perspective change, it appears the same to both. This leads to a set of assumptions that each person can make about the other, such as “they can see when I fidget”, or “they can see that we are in a crowded place”. The fact that both people see the same thing is important because they both use their environment to make decisions on how they interact with the other person. Online, there isn’t much to see. So when you present an environment to a group of people, and you expect that they will use that environment to interact with each other, it seems prudent that you build in this “sameness of vision”

I like to consider that Erickson’s claim #1 is a component of the “social contract” that we each accept whenever we interact with one another. Part of that social contract is that I can see and interact with you only if I allow the same behavior from you. I can’t walk around the world completely invisible, and expect that I should be able to see everyone else, but not allow them to see me (this is a common mistake in online interaction systems). The social contract is about reciprocal relationships of information. I can have access to certain information only if I give up that information (or some similar information) to others.

Similarly, for claim #5, we must look at what people do with social information. Erickson says that “accurately presenting information is not the point of a social visualization; its primary role is to provide grist for inferences.” Consider again what people in a physical world environment do with social information. Walking into a room, its important to know only generally how many people are there (one, a few, a couple dozen, hundreds, thousands, etc.). Once we know generally how many people are around us, we change our interaction style to reflect this information.

There is a component of claim #5 that also includes claim #4. Erickson doesn’t speak about a user’s attention when using a Social Visualization, but the movement of attention from micro to macro and back is important. When you enter a room in the physical world, you take in a macro view of the number and type of people there. You then move into a micro view to determine if you know any of them. So in this example, ambiguity is useful for enabling us to gloss over details we don’t care about (like the specifics of exactly what all 212 people in a room are wearing), but being able to direct your attention to a specific micro view that enables you to interact with your 2 friends is similarly useful.