This dataset is provided as supporting information for the submitted paper: "Predicting power conversion efficiency of organic photovoltaics: models and data analysis". The dataset consists of two CSV files. The first CSV file contains data on the characteristics of organic solar cell donor candidates obtained from the Harvard Clean Energy Project Database (CEPDB). This includes SMILES strings, HOMO, LUMO, and Scharber Equation parameters derived by means of density functional theory. The second CSV file contains experimental data obtained from the Harvard Organic Photovoltaic 15 (HOPV15) dataset, including SMILES strings, solar cell architecture, HOMO, LUMO, optical and electronic gaps, power conversion efficiency, open circuit voltages, short circuit currents, and fill factors.