The rapid advancement of smart technologies is arguably widening the digital divide, making the lives of people without easy access to the Internet increasingly challenging. This digital divide can exacerbate the educational gap between communities with and without reliable digital infrastructures, as learning increasingly relies on smart digital technologies such as learning management systems, learning analytics and artificial intelligence tools. In this paper, we discuss a distributed cooperative learning analytics approach for developing communities without reliable Internet access based on our look into the cases of Tanzania. To enable management of digital learning contents, learning analytics, and interaction with AI agents based on slowly transmitted and shared learning data, we propose Distributed Cooperative Learning Environments for Development (DCL4D), which provides mechanisms for supporting teachers and learners in distributed environments by extending and integrating Delay-Tolerant Networking, Semi-supervised Federated Learning, and Progressive Visual Analytics techniques. This is a first step towards the provision of distributed cooperative learning analytics for developing communities and helps to identify key challenges to pave the way for smart learning support systems accessible to all.