A geological study requires time and data. Currently, about 40% of study duration is spent looking for information. If a user already knows what data he needs, the current search systems can usually provide it. So we believe that these 40% are not only devoted to searching what the user is already aware of, but most likely retrieving data which usefulness and/or existence is not necessarily known. As a result, our aim is to facilitate access to the later so as to reduce the 40% of the overall time spent looking for information. We assume that this information is accessible and available in databases. This will be achieved through the creation of a new recommendation system adapted to our constraints.