Introduction
Below are the featured data science and analytics projects of Brian T. McElroy, based in Tucson, Arizona, USA. Please see my LinkedIn for background information. I combine a career in computational neuroscience with creative technology in production environments. I've organized my portfolio below into sections so you can quickly find the type of projects that are relevant to you.
Featured DS/ML Skills
-
Tensorflow and keras
-
SQL / relational databases
-
Analytics and statistics
-
sklearn
-
Cluster Analysis
-
Computer Vision
Upcoming Projects
Below are some of the current projects that will be put into repositories in the near future:
-
University of Arizona's Hearing the Invisible is an art installation to explore the insights that turning brain waves into music can provide. Besides being a production manager and lighting designer for the project, I am also writing a patented protocol for translating EEG data into lighting DMX data. There are multiple difficulties faced in allowing for a visual representation of lighting data, primarily having to do with optimal refresh rate for visual acuity and intensity perception at high frequencies. Because DMX is sampled at 40 Hz and meaningful brain waves go up to 60 Hz or even 120 Hz depending on the region, resampling introduces challenges. It is likely that sampling z-scores of brain wave ranges (delta, gamma, etc.) will result in a far more visually acute effect.
-
A private medical education company is seeking to predict churn risk and other measures of student success or failure in its student population. It also seeks to aid admissions counselors in directing prospective students towards particular fields based on test scores, personality tests, etc. This is a commissioned work that will have to be carefully anonymized for publishing on my portfolio.
EDA Projects
Exploratory Data Analysis (EDA) is crucial in understanding the structure, patterns, and challenges of data before delving into modeling.
- EDA (Sales) -- InstaCart
- EDA (Sales) -- Ice
- EDA (Logistics) -- Zuber
- EDA (Sales and User Data) -- MegaLine
-
EDA and ML (Regression w/ Hyperparameter Tuning) -- InterConnect
(This project also includes extensive regression modeling.)
Classification ML Projects
These projects focus on classification models such as logistic regression and other algorithms to predict categorical outcomes.
Regression ML Projects
Below are projects demonstrating predictive modeling with continuous outcomes, such as profit or sales.
Other Specialized Projects
Some tasks involve time-series analysis, text vectorization, image processing, or other advanced techniques.