In a pandemic crisis, first responders and policy makers need to gain insight from the data and make evidence-based decisions. But how can we make sense out of our data? And how can this be communicated to a wide range of audiences?

Visual analytics comes to our rescue. But what is visual analytics? And how can we achieve it?

Cook and Thomas define visual analytics as “the science of analytical reasoning facilitated by interactive visual interfaces.”  The field of visual analytics combines strengths from visualisation, data analysis, knowledge discovery, data management, analytical reasoning, cognition, perception, and human-computer interaction aiming at knowledge discovery and insight gaining from large and complex data sets. The interactive nature of visual analytics allows for the integration of human judgement into the data analytics process through friendly interfaces.

In STAMINA, we develop a range of solutions to support preparedness and response to a pandemic crisis. Our aim is to equip decision makers with the appropriate tools to deal with complex decisions at operational, tactical and strategic level. An essential element in this effort is the ability to manage data and gain insight, make credible predictions, have a comprehensive situation awareness and take the public opinion into account. Additionally, all this should be communicated among a variety of teams and the public.

Towards this goal, we not only provide data visual analytics but also go a step further and provide predictive visual analytics as well as situation awareness visual analytics.

STAMINA visual analytics in action

Data visual analytics

Data management and harmonisation

STAMINA provides comprehensive data management services. Authorised users can upload, retrieve and query data via an intuitive user interface (see Figure 1). The data is annotated according to the STAMINA Common Semantic Data Model and adheres to a defined Data Access Policy.

Predictive visual analytics

Managing ICU bed capacity

CHARM (dynamiC Hospital wARd Management) is an ICU patient flow discrete-event simulation. It models ICU bed capacity in reconfigurable zones and caters for isolated wards for COVID-19. Its objective is to support ICU bed capacity management during surges of patients with infectious diseases and plan for emergency and elective operations during pandemic crises.

The CHARM dashboard (see Figure 2) provides a friendly interface where users can enter input parameters and run simulations for different ICU bed planning scenarios.

Predicting the temporal and spatial spread of COVID-19

FACS (Flu and Coronavirus Simulator) is an agent-based simulation that predicts the spread of COVID-19 in a local area (i.e., county, city, etc.). FACS considers preventive measures and immunity as well as the behaviour of individuals based on their demographics. It predicts cases per facility, hospitalisations and fatalities.

The FACS dashboard allows users to run simulations for predefined locations and analyse the results in a temporal and spatial manner.

Situation awareness visual analytics

Common operational picture

EMT (Emergency Mapping Tool) provides a common operational picture of the current situation in an emergency crisis, including pandemics. Colour-coded alerts are displayed on a map. Users can see more details by clicking on the map and select the information that they wish to display. A snapshot of the EMT dashboard is shown in Figure 4.

Capacity and resource management

CrisisHub Capacity and Resource Management Tool (CRMT) allows for monitoring resources availability and shortages in a single or multiple institutions. It supports end users in monitoring the availabilities and shortages of one or multiple institutions and assists in resource allocation planning. ​Figure 5 depicts the map overview of CRMT.

Web and Social Media Analytics (WSMA)

STAMINA’s Web and Social Media Analytics (WSMA) dashboard is a listening tool that monitors and analyses discourse on Twitter and Reddit. It allows users to filter content by keywords, location, time period and language and it performs sentiment and statistical analysis of the retrieved posts. Through line plots, bar charts and words co-occurrences tables, WSMA provides users with a snapshot of the volume, trends and subjects of discussion around the monitored topics. A screenshot of the WSMA dashboard is shown in Figure 6.