The hierarchial concept of land use planning becomes less relevant in a society with continuous interactions between stakeholders, researchers, planners and politicians. In this context, land use negotiation rather than land use planning appears to be the most appropriate concept. In the negotiation process, good quality data about the land is important as land properties are, obviously, key elements to be considered. Case studies at farm and regional level have been analysed to explore answers to a number of questions. How can soil data be presented most effectively? What are the research needs? How can the large existing body of data be mobilized most effectively? Studies on regional land use in Costa Rica used methods in a logical sequence including projections, explorations and predictions of land use patterns. The work involved upscaling of data, obtained at farm level, to the regional level. Work at farm level focussed on prototyping procedures in which farming systems were 'designed' by close interaction between farmers and scientists, including applications of precision agriculture. Soil data demands were analysed, emphasizing the effects of using data with different degrees of detail together with the application of pedotransfer functions which effectively transform existing data into parameters that are difficult or expensive to measure directly. This not only facilitated interactions with stakeholders but also with colleague scientists in interdisciplinary teams. In addition, use of Geographical Information Systems allowed visual presentations of alternative geographical land use patterns that were associated with various scenarios, thereby facilitating the interaction processes. A plea is made to increase interaction of stakeholders and researchers by considering research programmes as vehicles for joint learning.