We have applied a new, robust and unsupervised approach to data collection, sorting and analysis that provides fresh insights into the nature of single-molecule junctions. Automation of tunneling current-distance (I(s)) spectroscopy facilitates the collection of very large data sets (up to 100 000 traces for a single experiment), enabling comprehensive statistical interrogations with respect to underlying tunneling characteristics, noise and junction formation probability (JFP). We frequently observe unusual low-to-high through-molecule conductance features with increasing electrode separation, in addition to numerous other "plateau" shapes, which may be related to changes in interfacial or molecular bridge structure. Furthermore, for the first time we use the JFP to characterize the homogeneity of functionalized surfaces at the nanoscale.