Performing cartographic visual search experiments online: opportunities and challenges

Cartographic experimentation constitutes an essential process towards the examination of several map reading aspects. Among those, the evaluation of both the efficiency and the effectiveness of different types of cartographic products is undoubtedly of great significance. This statement can be directly proved considering the clear influence of the produced outcomes on the related theoretical framework (i.e. evaluation and ranking of the design variables) as well as on the practical issues connected to cartographic production (i.e. development of specific design rules). Different experimental techniques have been utilized for the study of map perception and cognition. Such techniques include simple (e.g. reaction time recordings) or more sophisticated (e.g. electroencephalogram) approaches. Over the last decades, the implementation of eye tracking experiments seems to have a great influence in cartographic community since it can provide objective and quantitative results regarding visual behavior. It can also reveal critical patterns of visual scanning during the observation of several cartographic stimuli by delivering qualitative visualizations of the collected gaze data. Nevertheless, eye tracking remains an expensive technology in general. Although low-cost, do-ityourself, and webcam-based methods and devices have been proposed in the recent literature, such solutions seem to be sensitive (in terms of data noise) and require the careful consideration of the experimental parameters, especially in the case of being used for scientific purposes. Hence, in practice, the existing eye tracking methods are not considered adequate enough for running remote (online) visual search experiments. Computer mouse tracking is one of the most promising methods that could be used for conducting visual search experiments in cartography and related fields (e.g. landscape perception). Several research studies indicate that mouse movements data are significantly correlated with the corresponding gaze data collected during an experimental trial. Taking into account that the method is based on participants’ reaction using mouse (typical) clicks, it can be considered quite suitable for the performance of cartographic visual search (task-oriented, top-down) experiments (e.g. experiments where participants are asked to locate a target-symbols among different distractor-symbols, or asked to enumerate specific symbols). At the same time, the aforementioned technique is much simpler than eye tracking since it does not require a separate calibration of each experimental participant. Recently, Krassanakis and Kesidis (2020) proposed a new MATLAB toolbox (called MatMouse) which provides specific functions in order to build, run and analyze mouse tracking experiments. More specifically, the toolbox supports the analysis of mouse movements data via specific metrics. Moreover, it gives the opportunity of (collected) data exploration using several visualization methods (Figure 1).