A cluster analysis approach to sampling domestic properties for sensor deployment

Sensors are an increasingly widespread tool for monitoring utility usage (e.g., electricity) and environmental data temperature). In large-scale projects, it is often impractical sometimes impossible to place sensors at all sites of interest, example due limited sensor numbers or access. We test whether cluster analysis can be used address this problem. create clusters potential using factors that may influence measurements. The provide groups similar each other, differ between groups. Sampling a few from group provides subset captures the diversity sites. approach with two types sensors: (gas water) outdoor environment. Using separate type, we characteristics up 298 sample across these representative coverage installations. verify installed as result sampling, well other measures available over one year. Results show vary clusters, thereby providing evidence cluster-based differences This novel methodology sampling It generalisable any situation in which influencing known. also discuss recommendations future sensor-based projects.