We get to do lots of cool things. The common thread is that we want to make things better in the places that we live, and are committed to using the best and most appropriate tools for getting there.
R-Center Pilot: Community-Academic Partnership to Assess Readiness for Implementation of a Trauma Resilience Intervention at 3 Atlanta- based Haitian Churches
Under the leadership of Dr. Gilberte Bastien at the Morehouse School of Medicine, and together with colleagues at Emory University, and the University of North Carolina–Charlotte, we are looking to determine the readiness of Haitian-Creole churches to implement mental health interventions. The hope is to build a better understanding of the factors and supports that can facilitate health and wellness in (a specific) immigrant population.
Analyzing RFP applications with Natural Language Processing
So much data is qualitative. But qualitative data is really tough to analyzing in community-based projects. As a consequence, people tend to avoid it, or just pick relevant quotes out at random.
Together with Dr. Victoria Scott at the University of North Carolina: Charlotte, we are looking at whether there are textual indicators that predict which way funding decisions go. This is a broad mandate, and we have many different methods and approaches to take: from simple bag-of-words approaches to more nuanced long short term memory recurrent neural networks.
Formative Evaluation of the New Jersey Health Initiative’s Small Communities’ Hyperlocal Data Collaboration
Oftentimes, it can be difficult to even start a community-based collaboration and improvement project. People might not know where to get data, how to interpret it, or how to act on it. That’s where this novel project comes in. Rather than make decisions based on RFPs, the team out of NJHI looks to work aside communities is South Jersey through the whole of the project; from concept, to application, to implementation, to evaluation.
Here’s a one-sheet I put together on how to think about data in collaborative settings.
Together with friends and colleagues at Virginia Tech and the American Institutes for Research, we are studying how readiness at multiple levels (individual, school, community) and across multiple roles (student, teacher, parent, administrator) impacts attitudes and beliefs about school safety. To do this, we have collected a lot of pre-COVID data that we have begun to work through.
One particular area is how we can extract readiness from text-based data. This could be done with a bag of words approach, but we are also using word vectorization method to look for semantic relationships between our big areas of interest.
Here’s as video made by the AIR folks on the project.
School and Community Readiness in Colorado
For the past four years, we have been investigating school and coalition-level readiness. Recently, we have begun consolidating that data to explore the appropriateness of modifying readiness assessment and how readiness clusters together within a network.