This project revolves around using sketch recognition (via drawing on a tablet) to classify emergent bilingual students' Spanish writing ability. In the pilot study, I collected data from 10 people who had varying levels of Spanish writing proficiency. From this, I engineered features which take advantage of a lack of confidence in writing in those who are not proficient in writing in Spanish. Then, I trained various supervised machine learning models to cross-validate the data. You can see my code here where I display all of the data, results, and implementation of each feature.

This project aims to use biometric data to predict stressful areas for walking around a Houston neighborhood. My responsibilities in this large (15+ people, multiple states, universities) project have been to help with feature engineering for the classifiers and preprocessing the data to be used in the classifiers. I created a script to partition tens of thousands of points into various location-based bins in order to predict if that location is a stressful location or not. I have also been tasked with going to the neighborhood events to recruit people to walk in our study. 

This project consists of a small team of three people where we are collecting sketch data from children ages 2-5 and building an assessment tool to predict deficiencies in their fine motor skills. We developed an application that acts as the data collection medium where the kids can draw on, then created another application to annotate and view the data. Then, we engineered various features and used historical feature sets to create models that predict the children's age and if they are at risk of a deficiency. My responsibilities in this project were to help design and create the data annotation tool and to collect data from the children in the field.