Informing the detection of flash drought events by mining and modeling media reports
PI: Antonia Hadjimichael
Level of effort appropriate for the proposed project: 25% of postdoctoral associate’s time for one year
Plan for funding the remainder of the researcher’s salary: Current DOE-funded project
A list of specific areas of computational and/or data science expertise: time series analysis, fundamentals of text mining, fundamentals of machine learning
Any other requirements or expectations of potential ICDS Junior Researchers: attend weekly group meetings, work with ICDS RISE engineers
A list of specific objectives for work supported by this call: develop a working prototype of text mining workflow for flash drought event detection; reproducible and transferable code repository of workflow
At least one medium to long-term goal: medium term goal to submit a publication on the tool in the Journal of Open Source Software (or similar); long term goal to acquire sustained funding for the effort from DOE Office of Science
Connection of the project to ICDS’s mission: This multi-disciplinary approach combines natural language processing and machine learning techniques with physical drought detection methods, and brings together data from news reports with remotely sensed and modeled land surface data. The project directly supports ICDS’s mission to advance computational and data science approaches to pressing societal challenges, by demonstrating the value of integrating diverse data sources for the detection of hydroclimatic hazards.
Team member’s recent and/or planned engagement with ICDS: PI Hadjimichael is a member of ICDS’s Cyberinfrastructure Faculty Advisory Committee (CiFAC). The project will also utilize existing RISE engineer support and ROAR Collab computational time in the PIs allocation.