Combining Participatory and ESM: A Hybrid Approach to Collecting Annotated Mobility Data

Collecting continual labeled activity data entails considerable effort from users to label a series of activity data. We propose Checkpoint-and-Remind (CAR), a hybrid approach that combines participatory (PART) and context-trigger ESM labeling (ESM). Checkpoint-and-Remind has the advantage of user control but reduces users' burden in recording activities. Meanwhile, it features a context-trigger mechanism of ESM as a backup to remind users of labeling. Our preliminary evaluation of CAR with nine participants, who collected and labeled their mobility activity data for 15 weekdays, showed that compared with PART and ESM, participants collected a larger amount of annotated mobility data using CAR. In addition, participants had a higher annotation rate when using CAR than when using ESM. Our results show that the hybrid approach that combines manual and automated recording is promising. Our future work is validating these results and measure more metrics related to compliance with more participants.