Personal
Taylor Swift Song Data Analysis
- Extracted track data from the Spotify API using Python to analyze key audio features (tempo, energy, and danceability).
- Cleaned and normalized raw data using Pandas; removed nulls and handled duplicates.
- Created Matplotlib visualizations to explore trends in Taylor Swift’s music over time.
Click this link to view the Taylor Swift Song Analysis GitHub repository for more details about this project.
Traffic Sign Recognition Using Machine Learning
- Developed a convolutional neural network in Python using Keras and TensorFlow to classify traffic signs from image data.
- Achieved 95.8% test accuracy across over 1,000 images, using image augmentation and dropout to improve generalization.
- Demonstrated real-time object classification performance in simulation and documented findings in a project report.
Professional
Quantifying Vehicle Steering Discrepancies
Contribution:
- Restructured and normalized 500+ JSON log files using Pandas to enable structured analysis of steering behavior across autonomous vehicle test data.
- Visualized vehicle trajectory and signal data with Matplotlib to identify erratic steering patterns from 50+ vehicle tests, contributing to feature root-cause analysis.
- Engineered features and trained a logistic regression model to detect steering anomalies, achieving 93.4% accuracy and improving early error detection.
Proactive Road Construction Detection System
Contribution:
- Scraped construction and road closure data from 48 Department of Transportation (DOT) websites using Python (BeautifulSoup) to anticipate road hazards and reduce reliance on evasive event logs.
- Cleaned and normalized raw HTML data with Pandas, ensuring consistent, high-quality inputs for system integration and automation.
- Built an automated notification pipeline to preemptively flag high-risk construction zones, enabling preemptive disabling of ADAS features and reducing customer-reported issues.