Junying Li Email & Phone Number

Los Angeles, California, United States

-Spearheaded comprehensive research on 1400+ skincare products by analyzing ingredient functionality, customer feedback, and profit margins, utilizing interactive data visualizations via Plotly to deliver data-driven insights-Engineered a Scrapy scraper using CSS selectors to efficiently retrieve functionalities of 1100+ ingredients-Executed feature engineering with Boolean-classification features to categorize ingredients andpercentile-based features to rank and normalize contributions, enabling comparative analysis-Accelerated the process of matching over 45k messy ingredients by 80% and doubled the accuracy of dataset alignment by leveraging TF-IDF vectorization and cosine similarity in the preprocessing pipeline

Ningbo, Zhejiang, China

-Improved a linear HAR model to perform rolling predictions for stock volatility that enhance precision by 25% -Developed and implemented a Random Forest Regressor model and an LSTM model, achieving a 30% increase in prediction accuracy for stock volatility -Conducted data preprocessing using a sliding window approach and performed hyperparameter tuning -Evaluated models’ performance via R2 and mean squared error metrics for future improvement

Los Angeles, California, United States

-Generated 3 box plots and line plots using Seaborn, visualizing Busquets' contributions to Barcelona 2022-23 season -Performed Principal Component Analysis (PCA) to reduce the number of metrics from 34 to 12, while capturing 95% of the cumulative explained variance -Devised a recommender algorithm utilizing Cosine Similarity to identify top 15 similar players to Busquets -Designed 6 radar plot posters to compare candidates' skills with Busquets and rated them via age, injury history and price

Los Angeles, California, United States

-Conducted in-depth analysis of home-field advantage in Premier League by engineering 5 new features -Applied 2-sample T-test with SciPy, verified assumptions of homoscedasticity with Bartlett Test and normality with Shapiro-Wilk Test, confirming the presence of home-field advantage with a p-value of 0.000381 -Evaluated significance of fan-presence: team performance deproved by 25% when playing in empty stadiums during Covid -Designed appealing visualizations using Canva, enhancing the readability and aesthetics of bar plot and line chart by 60%

San Francisco Bay Area

-Spearheaded 3 marketing campaigns that attracted 300+ clients and exceeded 46% conversion rate. -Amplified brand visibility on social media channels: drew 32721 viewers and gained 1600+ engagements, which successfully attracted 100+ new clients to purchase health insurance. -Drove a 166% increase in international health insurance sales at UCLA compared to the previous sales season. -Ranked #3 among 98 marketing interns and awarded the Best SMer due to outstanding sales performance.

Beijing, China

-Collected, identified, and analyzed 2126 differential gene expressions in COVID-19 patients using dplyr. -Visualized key findings through three key plots (volcano plot, GO enrichment analysis diagram, and KEGG enrichment diagram) to enhance data interpretability and effectively communicate complex genetic data.-Performed GO enrichment test to extract 24 clusters of lymphocyte signaling pathways using the “enrichplot” library.-Co-authored an 8-page research report and published on Francis Academic Press.