A minimal, hackable annotator for user studies. Originally created for SAMoSA (IMWUT/Ubicomp 2022).
Anytime you need to manually annotate user sensor data. For example, IMU data from smartwatches, smartphones, etc. This tool will log unix timestamps to a csv, which can later be synced to sensor data.
The tool randomly presents an activity and context, given a set of contexts with corresponding activities. First, a context is chosen at random, and then the trials of all the activities pertaining to that context are randomized. For example, in the workshop context, the 15 trials (5 activities × 3 repetitions per activity) are presented in a random order.
- Redo Trials - You (the experimenter) will inevitably make a mistake!
- Save progress - user study progress is stored in
study_data/pid/p_state.json, so you can restart whenever you want to! - Timer - in order to make sure you collect enough data! :D
Tested with python3.8
flask
python app.py
Make sure it's a number - this is used as the random seed!
Modify static/activity_list.json.