Measuring and reducing cognitive load in map-based data visualisations. An analysis of communication mechanisms to improve decision-making processes in three different contexts: climate, media and administration

Abstract

[eng] Maps are one of the most versatile visual representations, and a common metaphor present in magazines, reports, digital media, print and television. Their familiarity and popularity often make them prime candidates to act as a data representation framework for communicating social, climate, political and other data. Although the use of maps is often the most obvious choice when presenting data in a visual format, sometimes their use to support the visualization of complex data in infographics and visual tools is not the best alternative, especially when decision making is key. Through three different scenarios: weather, media and public administration, we will measure the impact on the cognitive load of our users by analyzing uncertainty, spatio-temporal and multi-categorical data respectively. We use different metrics, as well as qualitative and quantitative methodologies to analyze the effectiveness of geospatial visualizations in decision-making tools. We examine ways to reduce cognitive load to improve user experience and understanding of the data presented. In the first scenario (weather), we simplified visual information channels to shorten response time and understanding when visualizing uncertainty. In the second scenario (media), we use different maps depending on user objectives in order to strengthen trend detection over time. In the third scenario (management), we customize the visualizations, allowing the user to dive into more specific data to enhance the comparison between various categories. The common goal of these three scenarios is to support decision making. In our first experiment (reducing the cognitive load on the climate representation of uncertainty), the effects of simplifying the visual coding, in terms of results, time and user satisfaction, were clear. When redesigning or simplifying a visualization, we must take into account two aspects, the purpose of the visualization (exploratory, analytical, decision making, etc.) and the reduction or subtraction of visual information not relevant to the task. In addition, in establishing climate-related guidelines, the collaboration of subject matter experts, along with user experience, graphic design and visualization professionals, was key to achieving our goals. It was also crucial to take into account psychological aspects such as perception, cognition and cognitive load of our users in a given context of use. Another important aspect that emerged was the need (for climate centers and institutions) to disseminate results and best practices, in order to promote their use and encourage constant improvement. In the case of static maps, when it comes to communicating global patterns over time, none of the representations studied is recommended, as users completed the tasks without the certainty that their answers were correct, which affected decision making. Although some maps stand out over others in specific tasks, our recommendation is to find an alternative representation to the static map or to use it only in cases where the temporal pattern follows a clear spatial direction. Of note is the weight given by users to the familiarity of some maps and how it affects their perceived performance in completing tasks. Finally, the effects of personalization together with intuitive interaction design are key to the dosage of information in visual information architectures. The user has just enough information to draw conclusions or make decisions, reaching the necessary level of detail through visualization. The general approach of designing tools for the majority or guaranteeing a minimum of satisfaction to a percentage of users does not always fit. Often, the needs of the same user profile may change depending on the type of task or context. These specific situations may imply a radically different way of displaying visual information.

[spa] Los mapas son una de las representaciones visuales más versátiles y una metáfora habitual presente en revistas, informes, medios digitales, prensa y televisión. Su familiaridad y popularidad los convierten a menudo en los principales candidatos para actuar como marco de representación de datos complejos, bien sean de carácter social, climático o político, entre otros. Aunque el uso de mapas suele ser la opción más obvia a la hora de presentar datos de carácter geoespacial en un formato visual, no siempre su utilización es la mejor alternativa, especialmente cuando la toma de decisiones es clave. A través de tres escenarios bien diferenciados: clima, medios de comunicación y administración pública, mediremos y analizaremos el impacto en la carga cognitiva de nuestros usuarios trabajando con datos de incertidumbre, espacio-temporales y multi-categóricos. Utilizamos diferentes métricas, así como metodologías cualitativas y cuantitativas para analizar la eficacia de las visualizaciones geoespaciales en las herramientas de toma de decisiones. Examinamos formas de reducir la carga cognitiva para mejorar la experiencia del usuario y la comprensión de los datos presentados. En el primer escenario (clima), simplificamos los canales de información visual para acortar el tiempo de respuesta y la visualización de la incertidumbre. A su vez, identificamos una serie de buenas prácticas y barreras en la comunicación de datos climáticos al público no experto. En el segundo escenario (medios de comunicación) utilizamos diferentes mapas en función de los objetivos de los usuarios, con el fin de reforzar la detección de tendencias a lo largo del tiempo. En el tercer escenario (administración), personalizamos las visualizaciones, permitiendo al usuario bucear en datos más específicos para mejorar la comparación entre varias categorías. En este caso la visualización pasa a ser la estructura de navegación. El objetivo común de estos tres escenarios, es entender los pros y contras del uso de mapa en estos escenarios y reducir la carga cognitiva en nuestros usuarios, a din de favorecer la toma de decisiones.