This talk is in three parts. The first deals with an aspect of the Weka project that has received little attention, namely the use of machine learning in agricultural applications. I will outline our experiences in this field and present an application development framework which is a direct result of this activity. In particular, one project has met one of the challenges proposed by Kiri Wagstaff at ICML 2012. Second, I will talk about our work in data stream mining with a focus on classification within the Massive Online Analysis framework MOA. After a quick overview of what is in MOA I will present two recent results that indicate a need for caution and a statement of what constitutes state-of-the-art in data stream classification for practitioners. Finally, I will present the idea of experiment databases, a framework for machine learning experimentation that saves effort and offers opportunities for meta learning and hypothesis generation.