Home > General Meetings, projects > Datasets Paper meeting

Datasets Paper meeting

October 22nd, 2008 Leave a comment Go to comments

Met with Lorcan, Steve Knox, Erica and Susan about the datasets paper.

Discussed practicalities of recording and annotating data, especially the video – we decided that it would be worth having a camera on each participants desk, to aid later annotation.

Lorcan stressed that it would be us (not him) that will be annotating the data, and that in the placelab set this took 1.5 hours for every hour of data.

It is proposed that we will record data over at least a month, and will be recording ground truths with annotation of video, and by keeping a diary of actual activities.

The following sensors will be considered for recording:

  • Ubisense
  • Bluetooth Sightings
  • Video Streams of key places
  • Twitter
  • Google Calendar events
  • Facebook status
  • Weather sensors
  • Currency exchanges
  • IM Status
  • Pressure Pads
  • Last.fm song lists
  • Keyboard activity
  • Mouse activity
  • flikr feeds
  • other reliable web sensor data

It was made clear that the administrators of this project (i.e. those present) should not be doing this if the results of this are not going to be of use to them in their research, and on this basis I should consider what my PhD is going to be.

My (unsaid) thoughts at this point were that I would like be involved, but I can’t see how it will benefit me, and perhaps I should help out with the writing of sensors etc.

There was some discussion of database storage, and Lorcan felt it would be simple to implement the ontologies in a database structure, I have some doubts over how simple it will be, but felt it would be a more robust mechanism than using construct as a datastore!

Categories: General Meetings, projects
  1. No comments yet.
  1. No trackbacks yet.