Would you use shared home activity data? (CHI Workshop CFP)
27 Oct 2008 - 10:19am
We are organizing a one-day workshop at CHI'09 (April 4, in Boston)
focusing on the development of free, publicly accessible multimodal
datasets recorded in the homes of volunteer participants. Although
this work was initially motivated by the lack of high quality sensor
data on which to test activity recognition algorithms, we would like
to extend the usefulness of such resources to other disciplines such
as design and home ethnography.
Our lab is currently funded by the NSF to produce continuous (24/7)
data recordings from at least four homes for six months each. Data
will include anonymized audio and video recordings from most rooms, as
well as activations from discreetly deployed sensing devices (motion
sensors, RFID, environmental sensors, etc). We also have annotators
on staff to label a subset of the activities in the dataset.
That said, we need a reality check. Would you actually consider using
this sort of resource? We will provide online tools to view the
datasets and search for instances of specific activities (e.g.
exercising, watching TV) or for use of specific objects (e.g. toaster,
phone, paper towels). I could speculate as to how a design
professional might use such information, but would prefer to hear your
ideas. If you believe the limitations of such datasets (e.g.
participant selection, observer effects) would render them useless in
your field, I'd also like to hear your critique.
Any suggestions for how we might improve the project overall would, of
course, be welcome.
Feel free to discuss here. If you were planning to attend CHI and
wish to have your opinions heard at the workshop, more information is
available at the link below. We would like to have a least one voice
representing the designer's perspective.