Brandon Sams
This project aimed to examine how well machine learning and neural networks would perform when trying to detect if a number was prime. A binary representation of a number was fed in as input, and a model was trained to detect if that number was prime or not.
Image Processing project that aimed to build a model that could determine if a face in an image was wearing a mask, not wearing it, or wearing it incorrectly. This ternary classifer was built using tensorflow and keras.
EP Means Clustering of Flight Data
Implemented an algorithm that could cluster distributions of data, rather than basic data points. Relied on computing the distance between distributions, (Earth Mover's Distance), and centroid computation to cluster airline routes by duration.
Data visualization project, where Tableau, Python, and R were used to visually illustrate how safe airline travel is, and how much safer it has been getting.
Bluetooth sensors were used to track traffic as it flowed around Austin, TX. This project dug into that dataset, and modeled traffic flow during different times of the day, and on different days of the week. Data was anonymized.
First foray into using Tensorflow for natural language processing and interpretation. Comments were classified as being positive or negative, based on tfidf vectorizer function.
This project aimed to examine how crime rates in Portland, Oregon compare to the national crime rates. Data was examined over a 30 year period.
This was an open ended exploration into a dataset. Data was cleaned, sorted, and visualized. Exploratory Data Analysis is the first step towards a larger project, and that was accomplished here.
This analysis was an examination of the dog-owner culture in New York City. Attributes such as breed, pet name, size, and age were examined in this analysis to determine how dog-friendly the city truly is.
A command line utility was written to make requests to the OpenWeatherMap API, and parse the JSON response it returned. This was an early project to get the hang of getting data from an API.