Experimental data on thermophysical properties of solvent mixtures such as aqueous deep eutectic solvents (DES) are scattered in scientific publications. It is desirable to integrate thermophysical properties with parameters that describe a DES in a human and machine readable format. On the basis of the Chemical Markup Language (CML), a standardized exchange format for density, viscosity, conductivity, and water activity of solvent mixtures was established and applied to represent published data on choline chloride/glycerol/water mixtures. In total, 300 different data sets served as a basis for data analysis by machine learning. Gradient tree boosting (GB) was used to predict thermophysical properties from the collected parameters, resulting in an excellent correlation between predicted, experimental, and simulation data. The experimental viscosity data was modeled assuming an Arrhenius dependency on temperature and by determining two parameters (eta(0) and E-eta). Integration of experimental and simulation data into a standardized exchange format makes data findable, accessible, interoperable, and reusable and enables machine learning methods. To facilitate data exchange, we recommend researchers publish experimental and simulated data on thermophysical properties as a CML-formatted Supporting Information with an associated digital object identifier (DOI).