Conception of a market garden management application in partnership with Lauzelle farm

Even though market gardening is attracting increased attention from consumers and researchers, it is not always easy for gardeners to make ends meet. For the practitioners, not only is gardening a tough task by the huge work labour it imposes, but it also requires important efforts for its planning and organisation. The study of different gardening techniques and the associated profitability of vegetables draws much attention among some researchers from UCLouvain. To support these studies, the university has recently started a project on a farm in Louvain-la-Neuve. In this context, this master thesis aims to provide a software tool which could support both gardeners and researchers. The first step of our work consisted in meeting some gardeners in order to better identify the main daily challenges they face in their planning and crop management. Meanwhile, it was emphasized that a key issue hampering research was the difficulty to gather reliable data sets. Following these observations, this dissertation builds a web application offering, on the one hand, a planning tool for gardeners and, on the other, a data gathering platform for researchers. Our solution allows a gardener to create a virtual representation of his garden and to divide his land into parcels and beds. He can either create or use existing identity cards of vegetables. Each of these cards contains a roadmap of successive cultivation operations for the cropping of the concerned vegetables. Our tool provides reminders of these cultivation operations throughout the year and helps the gardener keeping track of his actions. Several software engineering techniques were used throughout the development of this solution. It is important that the developed tool can be easily reused and maintained in the future. By using Python as main language, we ensure that most of the code is understandable by many developers, but also by most researchers. Concerning robustness, a large bench of automated tests is set up, continuous integration is used through Travis and maintainability is guaranteed by CodeClimate. We have successfully developed an application that is now online and available for gardeners and researchers. To validate the usefulness and relevance of the developed application, we presented our solution to several gardeners. The development of this application showed the gardeners' interest in using technology to release them from a tedious planning. It also revealed the complexity of building a generic crop model.