Briggs's Site - Projects

(one of my small projects was creating this website 🙂)

Aggression Prediction ML model analysis

For this project I used post-hoc analysis tools to understand the inner workings of various machine learning models used for prediction of aggression among autistic patients.

Causal Discovery Application

In this project I completed as a ML researcher at INRAE, I developed a full-stack application for agriculture experts to interact with a Bayesian network visualization and ChatGPT-4 chat assistant to perform causal inference and estimation.

Applying Convolutional Neural Networks to Plant Disease Diagnosis

This project applies various convolutional neural network (CNN) architectures, such as Xception and EfficientNet, to classify 38 different plant-disease combinations, achieving 95% test accuracy.

Patient clinical notes clustering

This project applies Latent Semantic Analysis in combination with various embedding techniques, such as UMAP, t-SNE, and PCA, to perform clustering on 10,000 instances of physician notes.

Animal Death Calculator (mockup)

This project is a mock up of a tool I'd like to spend more time creating. This tool calculates the estimated amount of animal deaths per year based on different forms of meat consumption. I've input some very rough estimates as a start, and mostly focused on getting the tool working. I'd like to collaborate with an animal right organization to add this tool to their website as a means to inform meat consumers of the impact of their diet on the number of animals killed in a year. This preliminary implementation is rather sloppy, as I did it in a morning. However, I think it could be recreated as a very impactful tool, given that it's also backed up by robustly research statistics on the number of deaths caused by different forms of meat consumption.

Patient Walking Metadata at MedRhythms

As a software engineer intern at MedRhythms, I worked on several projects related to patient clinical data, including surfacing patient time zone and calculating standard deviation of step precision for analysis by company researchers. I was part of the 'AlgoRhythms' team that worked together to implement the backend logic in Python/Django. Of course, I cannot demonstrate any part of this work as it was proprietary, but I will note that I very much enjoyed working for a company with such an intriguing mission and committed group of employees. I learned a lot in just a summer!

Mapping Maine Healthcare Data

In this project, I worked within a team of grad students, collaborating with a Maine based cancer center. We developed an interactive map overlaying various health and demographic metrics relevant to the center's cancer research goals. We implemented the program in Python with the GeoPandas and Folium libraries. I signed an NDA for this work and therefore cannot share the code publicly.

Trail Recommender

In this project, I led a team of grad students to implement a trail sorting algorithm that accounts for a variety of user preference inputs, such as trail length, dog-friendliness, elevation gain, etc. We implemented the code in Python and leveraged the Gower Distance function to measure similarity to the set of preference inputs. We also took advantage of the Pandas and GeoPy libraries.