Computer Science Research Projects
Evaluating Graduation Rates Between Private and Public Universities
Using visualization and statistical methods, we explored variations in graduation rates between private and public universities. We used numerical prediction methods to answer questions such as “Can we predict graduation rates?” and “How does the graduation rate vary for each university?” We used different data mining models and algorithms to compare efficiency of the constructed numerical prediction models. We found that the graduation rate is higher for universities that have a higher enrollment rate.
Testing the Testers: A Comparison of Software Testing Tools
Quality assurance testing is a large (and resource-intensive) part of software engineering as a whole. This research focuses on investigating and comparing software testing tools in order to reduce development time for software engineers and testers, as well as to ensure that all software is as reliable and safe as possible for the end-user. This study compares tools for software code coverage, defect detection, and mutation testing
Worcester Election Results: Cleaning and Preprocessing
The focus of this project was cleaning and preprocessing Worcester elections PDF data files containing scanned images. We converted these into accessible files, such as Excel files, and restructured data for further analysis. We analyzed the data sets for missing attribute values and detected and fixed errors that occurred under file conversion. We then analyzed attributes for dependency, outliers, and relationship by running machine learning algorithms such as K-Means, GMM, and Correlation.
COVID-19 Pandemic Trends: Cases and Death Rates
We evaluated a global COVID-19 data set recording new cases, new deaths, total cases, and total deaths. We applied a visualization technique to draw comparisons between countries. The visualization techniques suggest the number of new/total deaths depends on the number of new/total cases when compared between countries. We constructed numerical prediction models for the number of deaths and compared them by countries.
Big Data Technology Integration
Businesses currently face the challenge of integrating technologies as well as solving cases involving technological processes. My study involves developing an architecture and process to solve business-use cases with a variety of technologies that range from one to 30 years old. By creating this process using Big Data, Mainframe, and other software programs, we are able to process data and solve our business case.
Bank Marketing Campaign: Predicting Customer Response
This project focuses on predicting the response of a customer to a bank’s marketing campaign. I am using a
data set provided by a Portuguese bank. Various data analysis and data-mining algorithms are used to uncover
interesting patterns. The data analysis and data mining are done using Sci-kit learn library, R, and Python.
The project results can be used to improve the success rate of future bank marketing campaigns.
