Jiajian Liu

Highly motivated and detailed-oriented candidate passionate about using data to improve business performance and customer experience. Skilled at leveraging data to develop actionable solutions to business challenges and utilizing data mining and data visualization to create meaningful insights. Excellent technical aptitude and knowledge of programming languages, data analytics and data visualization.

• DuPont Analysis Automation, (Python, Jupyter Notebook, Pandas, Numpy, Matplotlib, Yfinance), Performed data fetching through Yahoo Finance API acquiring financial statements for companies in S&P 500. Designed a program that automatically converts key components into graphs for analysis mainly by Return On Equity, Assets Efficiency, and Leverages. Selected multiple stocks that perform the best in different sub-sectors for the purpose of optimizing the current company's portfolio.
• Portfolio Optimization, (Python, Scipy, Matplotlib, Numpy, Pandas, Yfinance), Acquired 5 years of historical stock data for selected stocks using Yahoo Finance API. Generated 10000 simulations using Monte Carlo Simulation predicting returns and risk of stocks with randomized weight ratio. Optimized portfolios' performances by Maximum Sharpe Ratio, Minimum Variance, and Maximum Return in both long-only and long-short portfolios. Back-tested optimizations with 1 year of data and generated returns and risks under conditions to help boost the company's profit on the stock portfolio.
• 2022 NBA Draft Prediction, (Python, Google Colab, Pandas, Scikit-Learn, Seaborn), Lead a team of 3 in building machine learning models to predict 2022 NBA draft picks based on historical data of college players' statistics and draft picks in the last 12 years. Preprocessed data through data cleaning metrics of eliminating or replacing null values. Applied feature selection using Recursive Feature Elimination to select most-related features to draft picks. Trained and tuned Logistic Regression and Balanced Random Forest Classifier on this rare event to forecast the drafting conditions. Evaluated model performance via K-Fold cross-validation and analyzed the result in the form of probability to predict who will get drafted and the ranks.