Sharlin Sanjid

Sentiment Analysis of Online Reviews
• Built a sentiment classification pipeline using spaCy and BERT (Transformers), achieving 87% classification accuracy on Yelp review dataset
•Analyzed 5,000+ customer reviews to identify sentiment patterns tied to product star ratings.
• Compared models using precision, recall, and F1-score to determine optimal performance
Tools: Python, spaCy, HuggingFace Transformers, Jupyter Notebook

Time Series Analysis for Stock Price Forecasting

• Modeled and predicted volatility for S&P 500 and NASDAQ indices using ARCH, GARCH, and VAR over a 10-year financial dataset• Reduced forecast error by 12% using optimized model parameters validated with Mean Squared Error (MSE)

• Created visualizations to support investment insights and inter-market trend discovery

Tools: Python, NumPy, Statsmodels, Matplotlib, Pandas

Greenhouse Yield Analysis Using ANOVA and Regression

• Analyzed ecological field data to study how salinity gradients affect tree growth, focusing on Specific Leaf Area (SLA) as a key metric.

•Applied linear mixed-effects models to account for species- and site-level variability, enhancing model interpretability

• Managed non-linearity and heteroscedasticity using log-transformation and polynomial terms

• Delivered evidence-based insights on environmental stressors impacting coastal ecosystems

Tools: R, lme4, nlme, ggplot2, dplyr, Linear Mixed Modeling

Sales Performance Dashboard (SQL + Power BI)
• Designed and built an interactive dashboard in Power BI to analyze 10,000+ retail transactions by region, product, and sales rep
• Cleaned and transformed data using SQL queries to extract metrics such as sales growth, monthly trends, and profitability
• Enabled real-time filtering and drill-down views, enhancing executive decision-making and reporting speed
• Tools: SQL, Power BI, Excel

Powerlifting