COMPLETE: Spinal Surgery Predictions
Will a particular patient reap the best benefits from undergoing spinal surgery?
Selected machine learning algorithms and medical data are used to predict pain and disability outcomes for spinal surgery.
Legal Document Classification
Can we predict whether a case will be settled based on early phase legal documents?
This project uses deep learning with legal case files to predict settlement and other legal outcomes.
COMPLETE: Adult Despair Classification
Can we provide earlier interventions for desperate people by identifying them at a younger age?
In this project, we use mental health and demographic data to better identify people in need.
With the addition of a Chief Data Scientist, Senior Data Scientist, and Data Scientist we have been able to engage in a number of collaborative projects. Those projects are described in further detail below.
Adult Despair Classification – a project in collaboration with Medicine Health and Society, using mental health and demographic data to better the drivers of the rising numbers of deaths of despair (suicide, alcoholism, drug abuse).
Spinal Surgery Outcomes – a project in conjunction with the VUMC Orthopedic Surgery team that uses selected machine learning algorithms and medical data to predict pain and disability outcomes for spinal surgery.
Peabody/Tennessee Education Research Alliance (TERA) – building a predictive model to assist principals in Tennessee to reduce teacher turnover, and improve outcomes for grade-school students.
Slave Societies Digital Archive (SSDA) – Collaboration with the SSDA to use state-of-the-art Natural Language Processing to speed the exploration of the ancestries of enslaved people from scans of handwritten records from the 1500’s through the 1800’s. The Data Science team completed an engagement with SSDA to develop a Natural Language Processing solution to identify the names of enslaved individuals from translations of historical documents. The solution will be applied to 2,000 unique volumes dating from the sixteenth through twentieth centuries that document the lives of an estimated four to six million individuals.
Wond’ry Innovation Garage – provided guidance and consultation to a team of undergraduates and graduate students working with RGP Consulting.
Legal Document Classification - a project in conjunction with the Vanderbilt Law School, using deep learning with legal case files to predict settlement and other legal outcomes.
VUMC Children’s Hospital – build tools and predictive models to allow VUMC Children’s Hospital to plan for changing caseloads and demands for personal protective equipment, ventilators, and staff due to COVID-19 impacts, and to explore the effect of policy changes.
Furthermore, our Data Science Team has conducted 42 consultations in which they meet with faculty researchers and students across campus to assess their data science needs or to provide guidance on projects. Of these consultations 14 have resulted in further work with the data science team, or have been written into a grant. The team has provided 5 colloquia, and presented 21 workshops serving faculty, students, and staff across Vanderbilt.