LaTiS: Enabling Scientific Data Interoperability via a Functional Data Model

The diversity of scientific datasets has been a hurdle for data analysis and scientific studies. Many data users spend inordinate amounts of time gathering and reformatting data before they can use it. This not only hinders the progress of basic analysis, it is a barrier for data interoperability. Attempts at data model unification have provided some relief but they tend to suffer limited adoption because they are specific to a given scientific discipline. We present a Functional Data Model (FDM) that captures semantics of datasets at an appropriate level of abstraction to facilitate data interoperability and ease of access.

At the University of Colorado's Laboratory for Atmospheric and Space Physics (LASP), we have been developing the FDM for describing and manipulating scientific datasets. This data model builds on the mathematics of Relational Algebra and the Relational Data Model which is the basis of many database systems. By specifying a *relation* to be a *functional* relationship between independent and dependent variables, the FDM can better capture the functional relationships that are inherent in scientific datasets. The FDM also maps well into the array-based data models that are common in the scientific community.

LaTiS is a software package that implements the FDM. It enables readers, operations, and writers that can be composed to orchestrate a data pipeline. LaTiS provides a RESTful service interface that has been deployed operationally at LASP to provide datasets to numerous applications in support of many NASA missions and beyond.

This presentation will describe the FDM and how LaTiS uses it to enable unified data access and server side operations for the LASP Interactive Solar Irradiance Data Center (LISIRD) and how it is being leveraged to provide an implementation of the Heliophysics Application Programmer's Interface (HAPI).