Demand Planning Analyst

Matthew Yeo

912-***-**** *************@*****.***

https://www.linkedin.com/in/matthew-yeo-8670aa265/ EDUCATION

The University of South Carolina, Columbia August 2020 - May 2024 Bachelor of Science, Major in Statistics

Relevant Coursework: Probability, Mathematical Statistics, Forecasting and Time Series, Applied Stochastic Processes, Applied Multivariate Statistics and Data Mining, Computing in Statistics, Big Data Analytics

WORK EXPERIENCE

Demand Planning Analyst Rather Outdoors May 2024 - Present

Developed and maintained accurate demand forecasts for $410,000,000 products, achieving a varying 90-95% accuracy.

Led cross-functional collaboration to refine forecast models, resulting in improved forecast accuracy and alignment with business objectives

Responsible for bi-weekly forecast updates of 3500 SKUs for Amazon’s orders

Maintained an accurate POS forecast with an 85-90% accuracy using Machine learning tools with Python

Built meeting slides that help sales reps compare POS, Shipment, and WOS to current and historical inventory

Partnered daily with the Sales team to align demand forecasts with customer insights and buying trends PROJECTS

League of Legends Player Performance Analysis R, Statistical Modeling, Excel

Analyzed 62 professional players across 25 variables (e.g., GD10, CSD10, DMG%) from the 2022 LCK Summer season to identify undervalued talent and win-rate drivers.

Isolated GD10 (gold differential at 10 minutes) as the strongest predictor of win rate (weak but significant correlation), enabling targeted coaching strategies for early-game leads.

Visualized outliers via star plots and chi-plots, revealing 2 teams (KT Rolster, Kwangdong Freecs) with players overperforming relative to team results. Forest Fire Risk Prediction Model R, Statistical Analysis, Regression Modeling

Regression tree model achieved 92% accuracy in classifying fire severity (none/moderate/severe) using temperature/humidity splits.

Discovered rainfall reduced fire likelihood by 75% (only 2 fires occurred in 8 rain events).

Insights could optimize wildfire prevention resources by prioritizing high-risk conditions (e.g., low humidity + high wind).

Identified temperature 19.85 C and relative humidity 42.8% as critical thresholds for severe fires

(p < 0.05).

Valorant In-Game Economy Forecasting R, Time Series Analysis (ARIMA), Game Analytics

Developed an ARIMA model to predict enemy team loadout costs (forecasting 10 rounds) and analyze the impact on match outcomes using 5,000+ round-level data points.

Identified economic tipping points: Teams with low economies (<5,000 credits) won 27% more rounds than expected, revealing strategic gaps in high-economy play. Earthquake Forecasting with ARIMA Modeling R, Time Series Analysis, Geospatial Analytics

Analyzed 100 years of seismic data (1900–1998) from the National Earthquake Information Center to forecast earthquakes 7.0 magnitude using ARIMA models.

Achieved 1.5 mean absolute error in 10-year forecasts (vs. actual 1999–2008 data), though model accuracy was limited by non-normal distribution (Q-Q plot).

Predicted rising earthquake frequency (19.5/year by 2008, 80% CI) based on historical trends. IT SKILLS - Microsoft Office Suite, R, SQL, Tableau, PowerBi ACTIVITIES - Intramural Volleyball, Collegiate Valorant Team, Student Associations(Taiwanese, Vietnamese, Filipino, Korean)

SKILLS - Organized, Diligent, Creative, Professional, Multitasking, Excellent Communicator, Team player

Contact this candidate