Engineering writing and communication is increasingly required of engineering undergraduate and graduate students; yet this call is rarely met through research-based methods. This research posits that deep investigation of the cognitive processes within engineering writing requires a novel method for data time-dependent analysis. However, no such method exists. This paper details our decision-making processes as we examine the available methods to analyze screen-capture video data of real time writing in an authentic engineering writing context. We detail the process of selecting the most appropriate approach and then provide a proof of concept for this method though analysis of screen capture video data of authentic engineering writing spanning several weeks. Complex issues in engineering education require creative interpretations of what can be used as "data," and similarly, creative and rigorous solutions for analysis. The results of this paper directly affect researchers doing real-time analysis of video data in all respects, and will also help other engineering education researchers to analyze non-traditional data.