Information visualization

Any activity carried out today – by companies, organizations or institutions – is synonymous with the production or collection of data to keep a record of this activity or those involved in it. The world of research is developing methods to make the most of this data, to mine, analyze and model it. The monitoring activity, which combines data collection and analysis, is at the heart of a process that enables socio-economic players to constantly adjust their position and strategy in relation to their environment.

The data collected and stored in this way is often complex, as are the phenomena from which it is derived, but also often because it is unstructured. Data mining algorithms are used to extract structural patterns or association rules from the data, the first form of knowledge gained from the data. Companies are interested in drawing up profiles of their customers to better target their sales campaigns, for example; observers of public debate will want to identify the players determining the barycentres of opinions. Because data is massive and complex, the results of data mining are often just as complex. These results need to be sorted and classified to arrive at a useful synthesis for decision-making, or to formulate an explanatory hypothesis for the phenomenon under study.

The success of the excavation operation depends on the ability to read the patterns discovered. The interpretation of these patterns, the evaluation of their relevance and the applicative potential of the results of the search remain a very human affair, the complexity and subtlety of which cannot today be entrusted to an automaton. In the end, only humans can judge the relevance of a result, appreciate its meaning and potential impact, and make the right decisions where necessary.

The field of information visualization is based on a fundamental observation: around 20% of our brain activity is devoted to processing visual signals ....