Visual programming languages, such as Pure Data (Pd) and Max/MSP, have been prevalent in computer music for nearly three decades. However, few shared and consistent research methods have emerged for studying the reproducible use of digital musical instrument (DMI) designers employing these languages. In this paper, we introduce straightforward methods for extracting design process data from Pd usage through automated version control and protocol-based annotation. This data enables visual and temporal analysis, which can reveal patterns of DMI design cognition and collaboration processes. Although our focus is on design, we believe that this approach could also benefit creativity studies and musicological analysis of the compositional process. We present the outcomes of a study involving four groups of DMI designers in a one-hour closed activity and demonstrate how these analysis methods can be used to gain additional insight by comparing them against participant survey data. In discussing how these methods could be enhanced and further developed, we address validity, scalability, replicability, and generalisability. Lastly, we examine motivations and challenges for DMI design cognition research.