Harnessing Data for Industrial Ecology Research: The Unified Materials Information System (UMIS) and the Yale Stocks and Flows (STAF) Database

A wealth of materials flow analysis (MFA) data has been gathered and analyzed since the inception of the industrial ecology field nearly three decades ago; however, a general and comprehensive method to collect and organize... [ view full abstract ]

A wealth of materials flow analysis (MFA) data has been gathered and analyzed since the inception of the industrial ecology field nearly three decades ago; however, a general and comprehensive method to collect and organize these data on the fundamental level of substances and distinct engineering materials (e.g., steel) is still lacking. The absence of such a method has led to repetition of work, poor reproducibility of results, and the segregation of research efforts that has diminished the overall impact of the industrial ecology research community.

The Center for Industrial Ecology at Yale University has played a significant role in generating and analyzing MFA data since the early 2000s. This data collection now spans 63 elements and several engineering materials across cities, countries, internationally, and globally, and over various time frames since before 1900 to the present day. However, these data are scattered across multiple non-uniformly formatted publications, reports, and spreadsheets. Therefore, consolidation of these data into a unified structure would be beneficial for their use in future research and to enable better cohesion of the industrial ecology research community.

Here, we present the novel unified material information system (UMIS) and its use in unifying Yale’s Stocks and Flows (STAF) data into a general, comprehensive, and standardized structure without loss of information. We then show how this transformed data is used to develop the Yale STAF Database, to store most of Yale’s material life cycle and criticality data.

The Yale STAF Database is being developed in MySQL, which is a highly documented, free, and open-source relational database management system, and also in collaboration with the U.S. Geological Survey, who will permanently archive and enhance it on the Internet in an openly-accessible format. The novel data structure upon which the Yale STAF Database is based, our UMIS, uses a system basis of a substance (e.g., Fe), reference timeframe (e.g., year 2000), and spatial boundary (e.g., United States). It includes domestic and trade flows, flows intersecting substance life cycles to enable analysis of the substance and material compositions of stocks and flows, and facilitates (dynamic) data analysis along a time series. The UMIS is compatible with input-output analysis and the compilation of life cycle inventories because it structures different types of data similarly (e.g., mass, money, energy) and is also consistent with the bipartite directed graph structure. This compatibility is borne out in matrix type UMIS diagrams that we develop and show here, which re-envisage the way that materials data are presented for complex data analysis tasks and complement existing visualizations (e.g., Sankey diagrams). Therefore, the Yale STAF Database and the UMIS are able to reconcile diverse industrial ecology datasets across its key techniques. These features are demonstrated here, showing how the Yale STAF Database and the UMIS can become foundational tools in industrial ecology research to advance our understanding and ability to quantitatively analyze substance and materials life cycles.