Analysis of three different experimental datasets (two published) in both R and Julia. The goals are (1) to demonstrate the reasoning and analysis of different types of experimental data; (2) to compare the results of Linear Mixed Effects models in R and Julia; (3) a tutorial on how to specify Linear Mixed Effects Models in Julia using the MixedModels pakcage.
Using data from Hillenbrand et al. 1995 on the acoustic measurements on American English vowels, supervised (Support Vector Machine) and unsupervised learning (k-means, Gaussian mixture models) algorithms are used to classify and create clusters, respectively. For supervised learning, the formants are used as predictors. For unsupervised learning, the formants and formant ratios are reduced to two principal components after normalization prior to model implementation. Cluster membership (but not confidence intervals) are visualized using R.
Retroactive simulations/analyses for the 2012 US Presidential election. Developed primarily from the ‘Desperately Seeking Silver’ homework and tutorial. Python (SciPy, NumPy, pandas) and Julia (DataFrames, GLM) used to create simulate outcomes. Working on extending to other methods and elections.
Parallel implementation of Google Python course implemented in base Julia. Devised as way to learn general Julia programming.
Penn Biotech Group, Fall 2014.
Management consulting project that involved developing a market entry strategy for a start-up. Performed primary and secondary research on market size, potential beneficiaries and under-/un-explored potential users of product. Conducted interviews with medical professionals, determined pricing strategies and financial modeling of market structure.