An image classification system built in collaboration with Dobot, Inc. that uses a convolutional neural net to predict what users are saving for based on images they upload to the Dobot app
SA Scorer is a web application built for the purpose of scoring raw data collected from spontaneous alternation using a 4-armed plus maze. Spontaneous alternation is a behavior task frequently utilized in behavioral neuroscience to test spatial learning and memory. The goal of this app is to streamline and eliminate human error in scoring. The application is built in Python with a flask framework and hosted on an AWS EC2 instance running apache2 and mod_wsgi.
A GUI to score and analyze Spontaneous Alternation data. This is the first iteration of and graphical user interface version of sascorer.com. Built with python, uses Pandas, xlrd, and guidata.
A couple of analyses I performed in Jupyter Notebooks for the lab I worked at. One is an analysis of ELISA protein assay pilot data. Another is an analysis of the indifference points of rats performing an operant chamber task for a pilot study.