Data Science Machine Learning

Siru Wu

Washington, DC ******@**********.*** 530-***-****

EDUCATION AND PROFESSIONAL DEVELOPMENT

University of California, Davis September 2020 – December 2023 Bachelor of Science (B.S.) in Statistics

Georgetown University August 2024 – May 2026

Master of Science (M.S.) in Data Science and Analytics SKILLS

Technical Skills: Python (Scikit-learn, Keras, Matplotlib, Pandas, NumPy, Statsmodels), R

(ARIMA, ggplot2), Excel (VLOOKUP, PivotTables)

Languages: English (Fluent), Chinese (Native), Japanese (Fluent), Spanish (Basic) INTERNSHIP

Deep Learning Intern

Wuhan Kedi Intelligent Environment Co., Ltd, Wuhan, China March 2024 – July 2024

• Developed a customized time series prediction model by integrating GRU into LSTM, using the Beihu Secondary Pump House flow dataset to improve forecasting accuracy.

• Implemented the data preprocessing workflow, including data cleaning, feature scaling using MinMaxScaler, normalization, and data splitting for data training and prediction.

• Optimized model structure and hyperparameters, achieving a 20% improvement in prediction accuracy, significantly enhancing model reliability and practicality.

• Visualized predicted and actual values with Matplotlib, effectively showcasing the model's precision and real-world applicability.

UNIVERSITY PROJECT

Health Risk Analysis Using Machine Learning

Georgetown University, DC August 2024 – December 2024

• Collected nearly 40 years of global chronic diseases data using the WHO API, addressing 50% missing data and standardizing formats. Retained 15,390 observations and selected 7 core health indicators for binary and multiclass model training.

• Applied K-Means, DBSCAN, and Hierarchical Clustering for unsupervised learning, reduced dimensions with PCA and t-SNE, and evaluated clustering using Elbow Method and Silhouette Score, optimizing performance by tuning parameters.

• Developed and optimized predictive models for binary and multiclass classification using Logistic Regression, Random Forest, and Gradient Boosting, improving performance through hyperparameter tuning and cross- validation.

• Built and deployed an interactive website using Quarto, with Plotly integrated for dynamic health risk visualizations. The website was developed with VS Code, featuring custom styling with CSS to enhance the user experience.

Predicting Washington, D.C. Precipitation Trends with LSTM and GRU Models Georgetown University, DC August 2024 – December 2024

• Retrieved and preprocessed 4,000 records of Washington, D.C.'s 2024 precipitation data using the API.

• Designed and implemented LSTM and GRU models, predicting daily precipitation levels, capturing 82% of trend variability, and enhancing precision for extreme weather events by 18%.

• Monitored and recorded carbon emissions during training using the CodeCarbon tool, achieving over 60% lower emissions compared to industry standards.

• Compiled an analysis report using Quarto and published it via GitHub Pages, featuring 7 technical modules and 40+ visualizations to present data processing, clustering, modeling, and results.

Contact this candidate