Hi there! Welcome to my homepage. My name is Sharon and I'm a data scientist with experince in data analysis and reporting, model development, database design, and data visualization.
- 🎓 A recent graduate from Master of Data Science in The University of British Columbia
- 💼 Worked as a Data Associate for the past 1+ years
- ✏️ Co-author of the
autopredictorPython package - 🗻 I enjoy small hike that leads me to scenic views
- 🐡 I love snorkeling
-
Language:
-
Libraries/Frameworks:
Robotic Arms Recognition Vision - Capstone Project
I employed computer vision algorithms, leveraging zero-shot learning transformers and fine-tuning Mask R-CNN (MRCNN), to enable robotic arms for precise object detection and dimensional measurement. The project culminated in an end-to-end pipeline where users can input an image into the trained model and receive accurate dimension outputs.
In this project, I utilized Docker and GitHub Actions to create a Python package called autopredictor. This package simplifies the machine learning model selection process for regression tasks by providing an intuitive and streamlined approach, significantly reducing the amount of coding required for model exploration.
Vancouver Crime Tracker Dashboard
An interactive dashboard using Dash to visualize 2023 crime data across Vancouver neighborhoods through geolocated maps, trend lines, and bar charts. The dashboard and app were created using Dash and hosted via Render, providing an accessible and dynamic way to explore crime trends in the city.
This project contains a deep learning project for waste classification using transfer learning. The model is trained on a labeled dataset of waste images and fine-tuned to achieve high accuracy in classifying different types of waste. The project includes scripts for training, evaluating, and making predictions with the model, along with a detailed Jupyter notebook that guides you through the entire process.
