MapReduce has become a popular programming paradigm for distributed processing platforms. It exposes an abstraction of two functions, map and reduce, which users can define to implement a myriad of operations. Once the two functions are defined, a MapReduce framework will automatically apply them in parallel to billions of records and over hundreds of machines. Users in different domains are adopting MapReduce as a simple solution for big data processing due to its flexibility and efficiency. This article explains the MapReduce programming paradigm, focusing on its applications in processing big spatial data. First, it gives a background on MapReduce as a programming paradigm and describes how a MapReduce framework executes it efficiently at scale. Then, it details the implementation of two fundamental spatial operations, namely, spatial range query and spatial join. Finally, it gives an overview of spatial indexing in MapReduce systems and how they can be combined with MapReduce processing.