Phineas Pham's Projects
Occupational Bias in Open-Source Pretrained Large Language Models: Analyzing Polarity towards Creative and Technical Professions
Analysis and prediction model of the bike sharing trend in Washington D.C. Area. Utilized advanced time-series techniques and R
Classify birds using CNNs pre-trained on chosen 10 pairs of birds
Optimize the childhood game of Chutes and Ladders using Dynamic Programming and Reinforcement Learning techniques
Used Python, Panda, and Numpy. Analyze relationship between Elon Musk's Tweets and crypo price trend.
This survey and analysis is used to discover how big the role of Slayter is when it comes to main meal of day. It is indespensible that Slayter is one of the most common space students go for food and beverage. However, in terms of main meals (which is, on average, students have three main meals a day), besides Slayter, Huffman, Curtis, Granville restaurants, and self-made food are also students’ go-to. After this survey, we hope to gain insight of students’ preference. With the result, we may can help develop Slayter to become a better market in a customer’s need-oriented way. The project is conducted with the help of Milo Dao and Minh Nguyen.
Fencing Video Review platform supports body-part segmentation for better analysis experience
Algorithm Bot helps you study Data Structure & Algorithms
K-Means vs KNN: A Performance Comparison on Image Classification of Bird Species
A study on Licking County Public Transportation, using data to propose suggestions on policy and improvements
A research review on Linear Programming and Simplex Algorithm
Magpie AI
With a movie summary, the model finds the most similar movie based on how similar their summaries are, by utilizing NLP techniques and Python libraries.
Gurobi Optimization Model to find the best route and means of transportation for a panoramic tour of Vietnam
My clone repository
A visualization and time series analysis of world space missions since 1957
Our data set is produced by the U.S Office of Personnel Management, providing statistical information about the Federal civilian workforce. This particular dataset is the newest quarterly update, consisting of U.S workforce data collected in June 2021. The purpose of this raw data set is to increase public access to high value, machine readable datasets, and they are accessible via https://www.opm.gov/data/. The original dataset contains more than two million observations and about 30 variables about each employee, such as their salary, length of service under the federal government, highest education level and so on. Approaching this dataset, our group would like to explore what the employment landscape is like under the U.S Federal government, as well as what factors affect the salary of an average Federal civilian employee. Particularly, we want to look at this dataset from the perspective of an undergraduate student looking for stable employment within the States. Hence, we have filtered the dataset we will work with to only contain information of employees posted within U.S territory and working full-time, which narrows our dataset down to just more than 30,000 observations that will be more substantial to our questions. Our approach to our first main question about the employment landscape under the U.S. federal government involve visually mapping out certain variables and how some of them might relate to each other. This is a broad, exploratory question for which we do not have any end hypothesis to test. From here, we could select variables of interest that might be more predictive of an average employee’s salary to put them in a multiple regression, therefore answering our second question about what factors determine salary for the U.S. Federal workforce.
YOLOv8 object detection algorithm and Streamlit framework for Real-Time Object Detection and tracking in video streams.