This diploma presents predictive markets, which are a means of aggregating information by using the underlying processes of the financial markets and offering forward contracts representing the probability of a proposal occurring. Publications in the field of finance, management or IT refer to the use of predictive contracts and demonstrate the good results achieved by this method. We wondered whether it was possible to design and use such a market in an academic environment, to predict the emergence of new theories and applications in the field of mobile communications. To support our thinking, we have put in place a prototype that should allow us to test the concept, study price calculation functions and information aggregation mechanisms. To begin with, we present the state of the art, building on the literature published since 1988 when the University of Iowa launched its Iowa Electronic Market (IEM) platform in the Presidential elections Bush-Dukakis. Since then, several researchers have looked at predictive markets as a forecasting tool in many areas: politics, sport, cinema, marketing, etc. We will compare predictive markets with other forecasting tools and then present in detail the predictive markets and a test to formalise their specifications. We will then illustrate this with examples from the main platforms currently in operation. A discussion followed on the fun orientation of the current platforms and a presentation of the uses that could be made of them and which would bring a real benefit. We then look at the analysis of the needs of both the investigator and the users. Looking for reasons for the complete absence of such platforms in research, we conducted a series of interviews with EPFL professors engaged in the MICS project. Their reluctance and concerns were taken into account when defining the prototype specifications. We use the (Spann and Skiera 2003) method to define the platform specifications. These specifications include the choice of the intended purpose, which in our case derives directly from our research proposal, the financial mechanisms underlying the transactions, and the incentives put in place to encourage users to use the platform and, above all, to make the most informed of them pass on their information to the market. These specifications lead to the presentation of four use cases. As our aim is to develop a prototype to test our assumptions, we are carrying out an overview of the available platforms and presenting the architecture of the latter, based on developments on the basis of USIFEX, which has not been in operation for four years. The prototype uses the flexibility of the subject programming based on the Python language in its version 2.3. We then work on the presentation of the prototype interface and interactions with users. It will then be necessary to refocus the interface to make it more user-friendly and enable us to guide the user during its sessions on the platform. The theoretical and software parts under roof still need to describe the contracts and how to formulate them. Once again, we are looking at the current markets to draw lessons that might be useful. Unfortunately, as no publications have been published, the definition of contracts is closely linked with the authors who draft them. We try to draw some lessons from contracts active in commodity markets to provide a framework for formalising the drafting of a contract. As this is only a proposal, it will still be necessary to test and complete this framework in order to ensure that the way contracts are drafted does not affect the relevance of predictions. Finally, we outline the work still to be done before the first real tests of the platform can be carried out. From the formulation of contracts to the study of combinatorial mechanisms for aggregating information, to pricing functions. We look back to the work carried out under this diploma and present some ideas for future research.