San Diego, California, United States
As a Research Assistant at the SDSU Machine Vision and Perception Lab, I contribute to the development of the Homeye platform, applying machine learning, computer vision, and advanced algorithms to provide comprehensive insights on over 15,000 properties in San Diego. My work involves automating data pipelines using web scraping, performing satellite and image-based environmental analysis, and developing convolutional neural networks (CNNs) for image classification, contributing to innovative solutions in real estate analytics.
• Developed an NLP-based school review sentiment analysis, uncovering key topics and sentiment trends in local school reviews.
• Analyzed Google Street View images for roadway segmentation and clustering, adding community rating indices based on infrastructure quality.
• Built a machine learning classifier to categorize house photos (e.g., kitchen, bedroom, backyard) and applied large vision models to generate automated text descriptions for enhanced property listings.
• Automated the retrieval of property pages from multiple real estate platforms (Zillow, Trulia, Realtor.com), expanding Homeye's data coverage.
• Integrated noise and radiation safety heatmaps into Homeye, visually mapping environmental impacts around properties.
• Conducted satellite image analysis to create a Greenness Index, enhancing community insights, and performed roadway analysis to estimate the effect on property prices.