Skip to content Skip to navigation

Learning Socio-organizational Network Structure in Buildings with Ambient Sensing Data

TitleLearning Socio-organizational Network Structure in Buildings with Ambient Sensing Data
Publication TypeTechnical Report
Year of Publication2020
AuthorsSonta, A, Jain, R
NumberTR242
PublisherCIFE
Place PublishedStanford University
KeywordsBuilding design, Design, sensing, Social Networks, statistical inference
Abstract

We develop a model that successfully learns social and organizational human network structure using ambient sensing data from distributed plug load energy sensors in commercial buildings. A key goal for the design and operation of commercial buildings is to support the success of organizations within them. In modern workspaces, a particularly important goal is collaboration, which relies on physical interactions among individuals. Learning the true socio-organizational relational ties among workers can therefore help managers of buildings and organizations make
decisions that improve collaboration. In this paper, we introduce the Interaction Model, a method for inferring human network structure that leverages data from distributed plug load energy sensors. In a case study, we benchmark our method against network data obtained through a survey and compare its performance to other data-driven tools. We find that unlike previous methods, our method infers a network that is correlated with the survey network to a statistically significant degree (graph correlation of 0.46, significant at the 0.01 confidence level). We additionally find that
our method requires only 10 weeks of sensing data, enabling dynamic network measurement. Learning human network structure through data-driven means can enable the design and operation of spaces that encourage, rather than inhibit, the success of organizations.

URLhttps://purl.stanford.edu/ss773tz5948
Citation Key2561