In recent years Visual Analytics (VA) has emerged as an efficient way of handling massive data sets, through combining human cognitive power with interactive visualisation techniques. In 2014, Defence R&D Canada noted that VA has gained significant momentum in defence, and this has been maintained, with VA technology already replacing traditional intelligence interpretation methods.
There are a number of key challenges to getting VA right in a defence context, however. These include being able to effectively and quickly deploy systems, being able to deal with large amounts of complex, high velocity, data and enabling user interaction in a way which converts this data into actionable intelligence.
First, for digital VA tools to work in defence they, and their supporting networks, must be deployable within austere environments. After all, the ‘office’ of an operational force could be anywhere from a megacity to a mountainside. At the same time, these systems must be rapidly deployable in situations where specialised personnel may not be available. This means that the infrastructure behind VA systems must be user-friendly and must allow end-users to manage and serve data themselves.
This data has always been complex – coordinating large bodies of military personnel whilst simultaneously predicting the actions of opposing forces naturally requires collating and analysing a massive amount of data. However, over the past decade this data has moved from being mostly static, fixed on a map or in a database, to being dynamic. This dynamic data may be of a novel type, such as multispectral video from RPAS or georeferenced social media data.This trend will continue as more sensors and data become available. In order to deal with the volume and velocity of this data analysis systems must be able to ingest data and make it available for analysis in real-time.
However, this is not enough to unlock the power of VA in a defence context because data on its own is not information and does not provide intelligence – it is just a lake of data. The user must be provided with a combination of different types of data be able to interact with it in order to extract insights. An example of this can be seen in a text analysis application created by Luciad and SAP. This application took twenty years of unstructured data from diplomatic cables, plotted events mentioned in these cables in time and space and provided users with multiple views of the data. This included a word cloud highlighting the most significant phrases and a chart showing the frequency of mentions over time. All elements of the application were interactive, allowing users to dynamically filter events by different terrorist groups (e.g. Boko Haram or Al-Shabaab) and draw out patterns of activity.
I see a bright future for advanced Visual Analytics in defence. From safeguarding critical assets to creating the digital infrastructure for future conflict, Visual Analytics will be at the heart of future defence systems, enabling more effective data analysis and enabling intelligence analysts to work faster and more effectively.