Abstract: Choice experiments are a type of survey which can be used to model user behaviour, in which survey respondents are presented with a choice of combinations of features and forced to one set of features. A computational system for performing choice modelling is described. This system comprises a user interface module, that allows an end users to specify what they set of features they wish to investigate (the attribute space), and constraints. A design generator module is used to generate an experimental design that matches the user defined attribute space. A range of different designs can be generated and compared to identify the optimum design. The design generator allows the user to specify their exact problem, rather than forcing them to modify their problem to match a known experimental design. An online survey assembly module is used to generate survey templates and a data collection and sampling uses the templates to display treatments to the respondents.