KVMesh is an example of a wide variety of mesh-based scientific data stored in key-value form suitable for processing by Big Data technologies such as Apache Spark or Hadoop...
The pictures here (click to enlarge them) illustrate that mesh-based scientific data has well a defined data model, is highly structured and often rich and diverse in logical, geometrical and mathematical relationships. This is in stark contrast to unstructured text and key-value pairs typical of Big Data workflows.