Focus on the important details, such as:

• Flows of essential information, expression, and control;
• The affordances that enable the user to affect the system, and how the system and the participants respond;
• How to make state visible audible, and tactile;
• Social aspects of the experience, and how your system affects them.

A suggested approach is to carry out the process described in [3]. Rehearse on each other before you bring in your subject/friends. Realize that lo-fi prototypes can be buggy! Work to identify and fix bugs in advance. Assign the suggested roles -- such as greeter, computer, observer -- to members in your group. A creative and constructive approach is to think of yourself as conducting a workshop [1, 2], in which participants explore important aspects of the experience of your installation, but without the technololgy. Boil it down to the essentails. Perform a "debriefing" at the end [3].

You may create whatever variations on this process you like, to the extent that you are very clear in your sense of what you're doing. Ask me (by email or in person) if you're unsure.

Your write-up could go 3 conference pages. One section should be data, not the raw data, but in an organized form. Focus on the relevant stuff. Then provide a section of analysis, followed by lessons learned, and next steps.