Group Detection Based on User-to-User Distance in Everyday Life for Office Lunch Group Recommendation

Companies have recently been introducing an event system called shuffle lunch, which aims to effectively utilize lunch breaks. The system has been garnering attention from many enterprises because it livens relationships and strengthens cooperation between departments and individual workers. In existing shuffle lunch systems, lunch groups are randomly generated. Random groups however, can cause problems. For example, some people might feel awkward about eating with new people, groups might not agree about when and where to eat, or some people may be absent. To form better lunch groups, the relationships between potential group members have to be known. In this study, we propose a method for dynamically detecting groups by using smartphones to measure the daily physical proximities of people. We also developed an Android application to realize our proposed method. We evaluated our system through a series of experiments and found that our proposed method can accurately detect groups, based on the proximities measured by the Android application.