In France and Europe, farmland represents a large fraction of land cover. The study and assessment of biodiversity in farmland is therefore a major challenge. To monitor biodiversity across wide areas, citizen science programs have demonstrated their effectiveness and relevance. The involvement of citizens in data collection offers a great opportunity to deploy extensive networks for biodiversity monitoring. But citizen science programs come with two issues: large amounts of data to manage and large numbers of participants with heterogeneous skills, needs and expectations about these data. In this article, we offer a solution to these issues, concretized by an information system. The study is based on a real life citizen science program tailored for farmers. This information system provides data and tools at several levels of complexity, to fit the needs and the skills of several users, from citizens with basic IT knowledge to scientists with strong statistical background. The proposed system is designed as follows. First, a data warehouse stores the data collected by citizens. This data warehouse is modelled depending on future data analysis. Secondly, associated with the data warehouse, a standard OLAP tool enables citizens and scientists to explore data. To complete the OLAP tool, we implement and compare four feature selection methods, in order to rank explanatory factors according to their relevance. Finally, for users with extended statistical skills, we use Generalized Linear Mixed Models to explore the temporal dynamics of invertebrate diversity in farmland ecosystems. The proposed system, a combination of business intelligence tools, data mining methods and advanced statistics, offers an example of complete exploitation of data by several user profiles. The proposition is supported by a real life citizen science program, and can be used as a guideline to design information systems in the same field.