Data Model
Structured, standardized, but flexible
HISTORICALLY
Historically, database systems were designed for banks, not for science! Relational databases have a rigid design, making adding/removing categories of information slow and difficult. What is needed is a data model that never turns away data.
Traditionally, materials data systems were primarily used to archive unstructured data (e.g. safety data sheets) for legal purposes, rather than using materials data for competitive advantage. A data model that makes curating data easy and actively rewards the effort with insight is needed.
The Future: Data Powering Competitive Advantage
By partnering with and learning from customers across different materials and chemicals classes, Citrine’s team of data engineering and material science experts have developed a materials data model, the Graphical Expression of Materials Data: GEMD.
FLEXIBLE, BUT ENFORCES STANDARDIZATION
The GEMD format is flexible enough to take in all data but enforces standardization where needed through templates and data bounds. This standardization provides consistency across an organization so that data can be discovered, aggregated, and analyzed.
CAPTURING THE FULL CONTEXT
GEMD captures the full context of a materials history. Data from procurement, processing, and characterization labs can be linked through unique identifiers. Both ideal processing conditions and real batch operating parameters can be recorded. This data is then unique to your business, AI-ready, and reusable in future projects.
ANALYSIS
Materials and chemicals processing history can be visualized in graphical format to understand the full context and linkage. Data can also be aggregated into tables for analysis, visualization of trends, and use in AI.
