Polymer nanocomposites, with the huge range of property sets (both structural and functional) they can exhibit, could pave the way towards “materials by design”—one of the promises of nanotechnology. Hindering the mass adoption of polymer nanocomposites is a limited understanding of the complex relationship between processing steps, structural parameters at the nanoscale, and resulting macroscale bulk properties. To this end, the work presented in this thesis reports on the investigations of four polymer nanocomposite case studies, each addressing effects of different variables, through NEXAFS spectroscopy—a synchrotron technique that offers rich chemical and structural information. In the first case study, non-covalent interactions resulting from electrospinning of a polymer blend of polydimethylsiloxane–poly (methyl metacrate) filled with carbon nanotubes are addressed, as well as effects of nanofiller diameter. The second study investigates the thermoactive behaviour of ethylene vinyl acetate j multiwall carbon nanotube composites through in situ temperature-resolved NEXAFS spectra, and an actuation mechanism is proposed. The third case study explores ageing effects of this thermoactive nanocomposite, and reports on the lifetime of non-covalent interactions. Finally, the fourth case study explores the effects of excessive sonication, which is seen here to drastically damage nanofiller and resulting matrix–filler interactions. This work represents four additional points in a growing dataset from other studies of polymer nanocomposites that—when sufficiently populated—could be mined through clustering algorithms and machine learning approaches to extract the elusive processing–structure–property link, which will enable the wide-spread exploitation of polymer nanocomposite technology.