ML&DS
NeoViews - Predicting Soccer Player Locations
-
Description: Developed a Long Short-Term Memory (LSTM) model to accurately predict soccer player locations throughout an entire match, achieving a prediction accuracy within 10 meters. Benchmarked the model against various optimization methods, including Bayesian optimization and linear regression, to evaluate performance improvements and accuracy.
-
Media:
Pathfinding Algorithms Visualization
- Description: Developed a visualization tool for Q-learning and Dijkstra's algorithms, focusing on practical applications of reinforcement learning in navigation and optimization tasks.
- Media:
HVAC System Optimization
- Description: Utilized ML and physics-based modeling to optimize energy use in air handling units (AHUs), integrating data science to enhance building sustainability. This includes leveraging an ensemble model (GBR, LSTM, Batch-RL, etc.) to create the best result for the complex system. Project is currently being evaluated in the physical HVAC system of a building.
- Media:
Laplacian Over Gaussian Operator
- Description: Developed a Python package for edge detection using the Laplacian of Gaussian (LoG) method to demonstrate image processing concepts with a packet of teaching materials to explain the mathematical basis of LoG. Includes customizable filters, example scripts, slides, a simple paper write up, a math proof, etc.
- Media:
Data Science Journal
- Description: Documented data visualization projects using tools such as Python (pandas, matplotlib, etc.), Excel, and Tableau, creating interactive network graphs, animated graphs, and more.
- Skills: Data Engineering, Data Visualization
- Link: Data Science Journal
The Phillipian State of The Academy
- Description: Led data analysis of a student survey with 80.5% participation, finding correlations using Python and libraries like pandas and numpy, and visualizing data.
- Skills: Data Engineering (Python), Data Visualization (HTML/CSS/JS)
- Link: Visit Project