This doctoral thesis is primarily devoted to miniaturized infrared spectroscopy, one of the most rapidly developing sub-domains in this very analytical discipline. A hardware modification of the worlds smallest near-infrared spectrometer is presented, greatly amplifying the signal-to-noise ratio and the reproducibility of the measurements. The performance of the modified spectrometer was rigorously assessed in a publication describing an EU relevant environmental application that involves liquid gasoline samples. However, the device is by no means limited to liquid samples since the versatile sampling accessory can also accommodate translucent and opaque solid samples. Taking the arcane character of near-infrared spectra into account, each of the reported analytical applications was tackled with the aid of multivariate data analysis. With sophisticated quantum chemical approaches, light could also be shed onto the particularities of molecular vibrations. Importantly though, such methods require substantial computational resources and hence it seemed rewarding to provide measures increasing their efficiency. In this context, three works have been published, describing straightforward techniques that reduce the computational burden of grid-based anharmonic calculations significantly. The synergistic capabilities of experimental infrared spectroscopy and quantum chemical methods were reason enough to conduct two collaborative works that clearly illustrate the importance of bilateral approaches in science.