The interpretation of complexity in isothermal calorimetric data is demanding. The observed power signal is a composite of the powers arising from each of the individual events occurring (which can involve physical, as well as chemical, change). The challenge, therefore, lies in deconvoluting the observed data into their component parts. Here, we discuss the potential use of chemometric analysis, because it offers the significant advantage of being model-free, using principal component analysis to deconvolute data. Using model data, we discovered that the software required a minimum trivariate data matrix to be constructed. Two variables, power and time, were available from the raw data. Selection of a third variable was more problematic, but it was found that by running multiple experiments the small variation in the number of moles of compound in each experiment was sufficient to allow a successful analysis. In general we noted that it required a minimum 2n + 2 repeat experiments to allow analysis (where n is the number of reaction processes). The data outputted from the chemometric software were of the form intensity (arbitrary units) versus time, reflecting the fact that the software was written for analysis of spectroscopic data. We provide a mathematical treatment of the data that allows recovery of both reaction enthalpy and rate constants. The study demonstrates that chemometric analysis is a promising approach for the interpretation of complex calorimetric data.