BIO-GEO ANALYTICS FOR DATA THAT IS SPARSE, HETEROGENEOUS, AND MULTI-DIMENSIONAL - WHAT IS NEEDED FROM NETWORK ANALYSIS? (Invited Presentation)

In the last 3-4 years, exploratory and predictive studies of mineral ecology and evolution have been advanced using visual analytics and network analyses. The science is interdisciplinary, the data multi-modal and the research questions have often been vague or ill-formed. Such undertakings become a team effort: a mixture of people with differing science and technical backgrounds, and the need for both in person and remote/ asynchronous collaboration. In exploring possible hypotheses, the reality of available data; quality, coverage, completeness and documentation confirm the data science 80-20 rule. 80% of the time you work with the data - in particular data structures - and 20% of the time is the science. Useful integration for scientific discovery requires close attention to those structures used in computing and analysis platforms in languages such as R and Python, and in technical environments. e.g. Jupyter notebooks.

What has become apparent during the (network, i.e. nodes and edges) analytics is that quantitative metrics focus only on the global or highly localized (at the node) aspects of the network. However, recent results for mineral and fossil networks clearly indicate nested structures and we currently lack the means to quanitfy these sub-clusters. This presentation will include the evidence for heterogeneous networks and prompt discussion of new network measures.