Analytical chemistry has never yielded such a wealth of experimental data as it does today, and this exponential trend shows no sign of abating. We continually advance the capabilities of our instruments and conceive innovative concepts, all in a concerted effort to naturally push the boundaries of our understanding regarding intricate sample matrices. Spectroscopic imaging, in the broadest sense, is certainly the field where we observe this acceleration even more pronouncedly. Analytical chemistry swiftly grasped the significance of processing acquired data for comprehensive exploration through utilization of chemometrics or machine learning tools. One can assert today that chemometrics undeniably constitutes an integral facet in the advancement of an analytical approach. However, we are now faced with a new challenge, as the experimental data accumulated for certain analytical techniques are so vast and massive that exploring them with such tools has become unfeasible, and this is by no means a computational capacity issue. Analytical chemistry is far from being the sole field affected by this issue, and one could argue that others have grappled with it long before us, such as, for instance, social media, to name just one. The purpose of this paper is to demonstrate that such a domain, which may initially seem distant from our concerns, can offer novel tools capable of overcoming these barriers, even though we are not necessarily dealing with the same objects. More specifically, we delve into the clustering of over 10 million LIBS spectra acquired as part of an imaging experiment aimed at exploring a singular rock sample. This will serve to demonstrate that an open-source library developed by Meta (formerly known as Facebook) can enable us to conduct a comprehensive exploration of this sample, a feat deemed impossible with conventional data analysis approaches.