Data for "The Coverage–Conflict Dilemma of VLM-Based Sensing in Smart Homes": Demand clusters with location and sensing-type labels
Authors/Creators
Description
This dataset accompanies the article "The Coverage–Conflict Dilemma of VLM-Based Sensing in Smart Homes," submitted to the International Journal of Human–Computer Interaction.
The data come from an online survey of 3,736 adults in their 20s through 60s in Japan. Free-text responses on task-execution demands (Q1) and information-provision demands (Q2) were translated into English and grouped into demand clusters. Each cluster was annotated with the home locations where the demand applies and the sensing types (Vision, Audio, other sensing) that could support it. The two files list all demand clusters and correspond to Tables S1 and S2 of the article's Supplemental Material.
Files
- S2_Table1_task_execution_clusters.csv: 75 clusters of task-execution demands (Table S1)
- S2_Table2_information_provision_clusters.csv: 93 clusters of information-provision demands (Table S2)
Columns
- cluster_id: cluster identifier (Q1-xx for task execution, Q2-xx for information provision)
- domain_category: domain category of the cluster
- cluster_label: short label summarizing the cluster
- unique_responses: number of unique responses used for clustering
- original_responses: number of original responses linked to the cluster
- representative_k_en, representative_k_ja (k = 1-3): three representative responses, given as the English normalization used for classification and clustering and as the Japanese original
- locations: home locations where the demand applies, separated by "; "
- standalone_candidates: sensing types that could each support the demand on their own
- required_combination: sensing types that must be used together ("---" if none). For a cluster with a required combination, standalone_candidates is "---".
The files are UTF-8 encoded CSV. Individual-level survey responses are not included to protect participants' privacy. The study was approved by the ethics review committee of Toyota Central R&D Labs., Inc. (approval number 24A-29), and participants gave informed consent.
During peer review, the files are accessible only through a private link. They will be made openly available when the article is published.