The legitimacy and effectiveness of current methods and theories that guide the construction of visualisations is in question and there is a lack of any scientific support for many of these methods. A review of existing visualisation techniques demonstrates some of the innate strengths and weaknesses within the approaches used. By focusing on the more specific task of developing visualisations for large sets of multivariate data, the lack of any kind of guidance in this development process is acknowledged. A prototype visualisation tool based on the well-documented techniques of Parallel Coordinates and Dynamic Queries has been developed taking into account these findings. Incorporating new and novel ideas addressing identified weaknesses in current visualisations, this prototype also provides the basis for demonstrating, testing and evaluating these concepts.