A predictive model to forecast customer adoption of rooftop solar

This study was designed to explore the overarching research question “How can we forecast residential solar rooftop adoption?” In order to answer this question, we utilized big data tools to forecast residential solar rooftop adoption likelihood based on both customer demographics and expenditure data. A model was developed utilizing SPSS, Azure, and ArcMap 10.3 software to yield an accurate prediction of household profiles likely to adopt solar rooftops. The study revealed a total of ten variables that significantly impact homeowners' decisions regarding whether or not to install solar rooftops. These ten variables accounted for around 23 or 24 percent of solar adoption variances. Importantly, this predictive analysis can be applied for all Energy Informatics use types; however, we focused on the residential parcel type located within LA County due largely to data availability. Future research is needed to expand on this analysis and to include the whole state of California, as to apply this analysis to other Energy Informatics cases, such as commercial properties and factories.