1. Defining your needs

Before jumping headlong into data collection, it's imperative to rationalize your needs. Whereas at the beginning of the 2000s, it was difficult to find digital data, today the massive movement towards open data, the widespread use of APIs (application programming interfaces) and transactional data are leading to the creation of such masses of data that the analyst must first and foremost clarify and formalize his or her information needs. This will have an impact on the choice of data to be exploited, the structuring model to be applied and the analyses to be carried out.

The analyst must define his or her needs as follows:

• What trade-offs/decisions will be made with the results of my datavisualization?

• What should my datavisualization represent, and if there are interactions between...