The smart building-integrated photovoltaic (SBIPV) systems have become the important source of electricity in recent years. However, many sociological and engineering challenges caused by temporal and spatial changes on demand-side and supply-side remain. In this paper, the barriers and traditional data utilization of SBIPV system causing the above challenges are summarized. Data-driven SBIPV was firstly proposed, including four aspects: Data Sensing, Data Analysis, Data-driven Prediction, and Data-driven Optimization. Data sensing goes beyond the technical limitations of a single measurement and can build the bridge between demand- and supply-side. Then, the demand-side response and electricity changes in supply-side under various environmental changes will also become clear by Data Analysis. Data-driven Prediction of load and electricity supply for the SBIPV is the basis of energy management. Data-driven Optimization is the combination of demand-side trading and disturbed system optimization in the field of engineering and sociology. Furthermore, the perspective of data-driven SBIPV, technologies and models, including all four data-driven features to make automated operational decisions on demand- and supply-side are also explored. The data -driven SBIPV system requiring much greater policy ambition and more effort from both supply and demand side, especially in the areas of data integration and the mitigation of SBIPV system.