Using data collected for the Cryptocurrency Sentiment Index project, I performed a linear regression analysis on Bitcoin sentiment and price data from a period of 31 days. A hypothesis test intended to prove the existence of a positive relationship between sentiment and price failed to yield a significant result. In order to see the results we were expecting, the linear model could be improved by increasing the number of observations or tuning our text classifier.
I created this project to visualize the relationship between a cryptocurrency's sentiment and its price. A set of Python scripts periodically collect tweets matching basic keywords about the top 20 cryptocurrencies by volume, and gets a rudimentary polarity score based on the contents of the Tweet.
With a web frontend built with React, users can view price and sentiment summaries for the top currencies and visualize the correlation of the two metrics on a chart. This application makes API calls to a simple Python application built with Flask.
As part of my experience in my Cloud Computing course, I worked on a team to deploy a private OpenStack cloud. On an existing OpenStack deployment, we created a virtual infrastructure on which to deploy our own cloud, consisting of multiple RHEL nodes, virtual networking, routing, and security. With the infrastructure in place, we used Docker with Kolla-Ansible to orchestrate our deployment.
In response to a community-request for album shuffle mode, I submitted a pull request to the Shuttle Github repo. This pull request was merged into the main development branch and the feature added the ability for users to shuffle whole albums rather than individual songs.
AUFoodTrax was an Android app for crowdsourcing reports of food truck locations on Auburn's campus. The app consisted of an interactive map for easy reporting and locating of trucks. Backing the app was a Spring Framework REST service and a MongoDB instance, both living in Microsoft Azure.