Modern approaches to analytics focus on the behaviour of the system and its properties, while very often the logical data model remains outside the scope of the analyst's attention. On large and complex projects, this can lead to disaster.
In the talk, I will explain why a good description of the data is more important than a good description of the behaviour of the system and give some practical advice on how to improve the quality of these descriptions. In addition, I will share a way to document data models using free software and Python. Therefore, the talk will be especially useful for those who have to pull off complex projects in the absence of requirements development systems.