I’m Jaewon, a UC Berkeley Data Science major with experience at DocuSign and MKS Instruments. My pipelines, models, and dashboards have powered decisions in fast-paced environments where speed and accuracy mean millions.
Personal, end-to-end projects in exploratory data analysis, forecasting, and data modeling/engineering.
Implemented LSTM and GRU deep learning models in TensorFlow to forecast Samsung stock prices, achieving R² = 0.95 with the GRU model; generated 10-day predictions providing actionable insights that advised against investment due to expected price decline.
Conducted an A/B test with Two-Sample Z-Test for proportions and demonstrated that the redesigned page did not lead to a higher conversion rate compared to the original.
Performed EDA and built a random forest regression model to predict housing prices, applying data preprocessing, model evaluation, and hyperparameter tuning; achieved R² = 0.80 and accurately predicted the target house price.
Employed PostgreSQL and advanced SQL with CTEs to perform multivariate analysis of COVID-19 data, identifying the most infectious countries and calculating a −0.751 correlation between GDP and infection rate, revealing GDP’s strong inverse impact on pandemic spread.
Built a PostgreSQL pipeline to identify the highest-paying Data Scientist jobs and the skills behind them, balancing frequency with pay to score “optimal” skills for upskilling.
Engineered a reproducible SQL cleaning pipeline for Nashville housing—de-duplicated, standardized fields, and enforced types—to produce analysis-ready tables with clear docs.
BeautifulSoup + Requests scraper that captures Amazon product title and average rating with proper headers, writes to CSV, and runs on a user-set schedule (e.g., every 12 hours) to track rating trends over time.
Personal dashboards built to explore real datasets and deliver insights.
Python pandas used for data prep and a multi-view dashboard presented in Tableau.
600 responses summarized with interactive visuals in Microsoft Power BI.
Geographical cost analysis with multiple breakdowns for Seattle Airbnb data.
Cleaned worldwide bike sales data in Excel and showcased with a dashboard.
I’m open to full-time roles in data science/analytics. If you're hiring or open to a conversation, feel free to connect or reach out to me!
Email: jaewon.shim@berkeley.edu
Phone: (510) 833-0481