fault, road cut, pavement

The importance of openly available digital resources

The recently published review by Kodaira et al., Reflections on solid Earth research highlights the importance of having digital resources such as EarthChem. Their work emphasizes the value of collaboration across broad research fields and the need for making data persistently discoverable:

Databases comprising a substantial number of data points describing the geochemistry of rocks formed through Earth history, such as EarthChem, or the records of sedimentation recovered from ocean drilling since the 1960s, such as through the International Ocean Discovery Program. These digital resources are providing access to previously undiscoverable data, building a foundation for multidisciplinary research and opening the geosciences up to powerful data science and machine learning approaches.

Logistic Regression Confusion Matrix

EarthChem at AGU 2020

Members of the EarthChem team presented multiple posters at the AGU Fall Meeting 2020.

An exciting Machine Learning/ AI experiment was presented by Arokium et al. as part of the CCNY Computer Science Masters Program. The team developed a logistic regression algorithm that successfully predicted whether two samples from PetDB are identical 86% of the time. The algorithm will be used to improve PetDB data entry and sample search capabilities.

EarthChem Communities were highlighted in a poster by Profeta et al., requesting feedback on how the future Volcanology Community should be set up.

EarthChem was also present at the ESIP Data Help Desk, providing video tutorials on the overview of our systems and why researchers should share their data through the EarthChem Library repository.