Lan Yu Email & Phone Number

Newark, De, Us

Newark, Delaware

• Spearheaded the development of a real-time outage detection and prediction system, integrating 10-seconds interval time series RGB image data (>100TB) with an on/off detection algorithm (utilizing Gaussian Process Regression (GPR) and Z-test), saving millions of dollars of smart meter installation in Freetown, Sierra Leone• Led the development of Supervised Random Forest and 1-D Convolutional Neural Network (CNN) models to classify 6 different lighting technologies (e.g., LED and Incandescent Lamp) in two images, achieving 98% testing accuracy• Applied multilinear regression and Markov Chain Monte Carlo (MCMC) sampling to model lighting spectrum and observed a 20% improvement in lighting technology changeover rate from 2015 to 2018• Utilized A/B testing to assess the effectiveness of the NYC lights-out policy, examining the correlation between lighting brightness (derived from 10-second image data) and the density of migratory bird populations (gathered through radar data)• Contributed to graduate teaching by providing peer-mentoring to a data science student, delivering lectures to 23 fourth-year undergraduate students, and serving as a Teaching Assistant for a graduate-level data science course

Newark, Delaware, United States

• Prototyped a K-Means User Segmentation Model in AWS SageMaker to classify 886,985 time-series risk-score curves, revealing critical pre-charge-off behaviors and patterns, which paved the way for advanced user segmentation and risk analysis• Calculated coefficients between cluster centers and time-series curves, establishing a benchmark for the User Segmentation Model, which involved 16,173 users• Extracted and transformed time-series data from Snowflake using SQL to generate a monthly private credit risk summary, while identifying anomalies and addressing coding issues concurrently• Led the refinement and application of the Life Time Value Model (KNN), addressed ad-hoc requests, performed ongoing testing and monitoring, and supported the marketing team in formulating strategic recommendations such as optimized paid search strategies• Collaborated with the project manager to create the Returned Customer Model for the marketing team, involving the selection of independent variables, definition of variables, and data cleaning

Shanghai City, China

• Designed and launched a registration management system and an online urban planning education feature, achieving 157,703 page views at the first year (top 3% among all International Knowledge Centers)• Supported the launch of a static city evaluation model to a dynamic city evaluation model (CityIQ) for 300 cities• Collected and updated 500 cities’ data and 128 best practice (50% increase) for iCity’s database• Designed and managed user acquisition programs (e.g., The parallel session: Artificial Intelligence and Future of Urban Spatial Planning of the United Nations World Geospatial Information Congress)

Pittsburgh, Pa

• Created and visualize American RIGO (Regional Intergovernmental Organization) and MSA (Metropolitan statistical area) by using Arc GIS software • Implemented web-based research on specific RIGO’s and to summarize that research into score sheets

Pittsburgh, Pa

• Collaborated with data collection of East Asian Library digitization projects• Helped with the development of Korean LibGuide (http://pitt.libguides.com/c.php?g=242421)• Assisted to general customer service in person and via telephone• Translated emails, documents, and videos from Chinese to English• Organized library orientation to international students, especially Chinese and Korean students

Doctor Of Philosophy - Phd, Energy, Environment, And Natural Resources Law (Data Science) - Stem Major

Master'S Degree, Policy Research & Analysis, 3.75/4.0