Skip to end of metadata
Go to start of metadata

A huge thanks to the three CDI Project teams who presented at our April Monthly Meeting.

An Interactive Web-based Tool for Anticipating Long-term Drought Risk

Caitlin Andrews Andrews, Caitlin Marie , a landscape ecologist in the Southwest Biological Science Center, explained how she used Rshiny and Amazon Web Services to create an interactive, online, front-end for a proven model of ecosystem water balance, SOILWAT2. This tool helps to predict and understand site-specific risk of future drought. Lots of lessons here for people who want to make user-friendly online tools out of more traditional scientific models within the USGS IT ecosystem. Code repository at https://github.com/DrylandEcology

Knowledge Extraction Algorithms (KEA): Turning Literature Into Data 

Matt Neilson Neilson, Matthew E. , a fishery biologist and co-lead for the Nonindigenous Aquatic Species Database program, delivered the line of the day: We are living in a machine-readable world. His project uses natural language processing and the xDD (eXtract Dark Data, formerly GeoDeepDive) literature database to improve, modernize, and greatly increase the efficiency of literature review. For people who used to walk to the library and photocopy stuff (and record radio songs on cassettes and dial with rotary phones), this is strange, but I will attempt to evolve with the times. See more information, like code repositories, in the Related External Resources links on the project's ScienceBase page.

Mapping land-use, hazard vulnerability and habitat suitability using deep neural networks

Jon Warrick Warrick, Jonathan , research geologist in the Coastal/Marine Hazards and Resources Program described the software tools, resources, and training workshops developed to allow USGS scientists to apply deep learning to remotely sensed imagery and better understand natural hazards and habitats. The 2 in-person workshops on these tools held in 2018 were able to accommodate only a fraction of the interested applicants. The CDI hopes to be able to provide more trainings like this to help build deep learning expertise and capacity in the USGS. See more at https://github.com/dbuscombe-usgs/cdi_dl_workshop and https://github.com/dbuscombe-usgs/dl_tools.


Log in to see the meeting recording and slides at the meeting page.

--
More CDI Blog posts

  • No labels