Identifying effective interpretation methods for magnetic data by profiling and analyzing human data interactions

Geoscientific data interpretation is a highly subjective and complex task because human intuition and biases play a significant role. Based on these interpretations, however, the mining and petroleum industries make decisions with paramount financial and environmental implications. To improve the accuracy and efficacy of these interpretations, it is important to better understand the interpretation process and the impact of different interpretation techniques, including data processing and display methods. As a first step toward this goal, we aim to quantitatively analyze the variability in geophysical data interpretation between and within individuals. We carried out an experiment to analyze how individuals interact with magnetic data during the process of identifying prescribed targets. Participants performed two target spotting exercises where the same magnetic image was presented at different orientations. The task was to identify the magnetic response from porphyrystyle intrusive systems. The experiment involved analyzing the data observation pattern during the interpretation process using an eye tracker system that captures the interpreter’s eye gaze motion and the target-spotting performance. The time at which targets were identified was also recorded. Fourteen participants with varying degrees of experience and expertise participated in this study. The results show inconsistencies within and between the interpreters in target-spotting performance. The results show a correlation between a systematic data observation pattern and target-spotting performance. Improved target-spotting performance was obtained when the magnetic image was observed from multiple orientations. These findings will help to identify and quantify the effective interpretation practices, which can provide a roadmap for the training of geoscientific data interpreters and contribute toward the understanding of the uncertainties in the data interpretation