Yihang Wang

• Deep-dived into the dynamics of US and Chinese stock markets, achieving an impressive profit of $13,000 through strategic investments.
• Diligently conducted market condition assessments every hour to exploit over 200+ hedge and arbitrage opportunities annually.
• Delivered in-depth analyses of market trends, synthesizing data from 50+ sources, generating comprehensive reports, and providing critical evaluations.
• Developed and implemented 10+ customized hedge and trade strategies, significantly enhancing specific trade outcomes.
• Showcased proficiency in using TradingView, analyzing 20+ indicators and conducting strategy tests for informed decision making.
• Capitalized on the Bitcoin market in 2018, demonstrating exceptional foresight and securing profitable positions.

• Conducted an in-depth analysis and predictive modeling project examining 144 NBA games over three seasons to predict match point distribution and team win probability.
• A stratified random sampling method was used to select games for analysis, ensuring full representation of NBA teams and seasons.
• Use advanced statistical software tools such as Excel and Minitab to collect and process in-game performance data with a focus on accuracy and data integrity.
• Apply least squares and logistic regression analysis to extract important predictors from multi-faceted game statistics and optimize prediction models.
• Identified seven key game statistics that have a significant impact on model predictions: field goal percentage (FGS), three-point field goal percentage (3PS), free throw percentage (FTS), offensive rebounds (ORS), assists (ASTS) ), number of turnovers (TOS) and number of free throws (FTAS).
• The fine-tuned forecast model has an adjusted R-squared value of 0.9145, indicating that the model is able to explain 91% of the spread variance.

• Effectively managed 10+ data cleaning, pre-processing, and visualization tasks utilizing Python and Tableau.
• Successfully migrated complex Excel models to Python, improving processing time by 50% and enhancing data management efficiency.
• Developed a predictive Python Scikit-learn model for the Australian Football League (AFL) as a university assignment, processing 100,0+ data points to predict high-scoring potential players.
• Performed meticulous pre-processing on the dataset, cleaning and normalizing data while reducing dimensionality by 30%.
• Partitioned the dataset into test and training sets, achieving a compelling cross-validation score of 0.88 out of a perfect 1.0.