Consumer Insights & Market Strategy for Beats by dre

I conducted a full consumer research analysis for Beats by Dre, turning survey data and qualitative feedback into clear strategic recommendations. By identifying key customer priorities, such as sound quality, battery life, portability, and smart-assistant features, I developed data-driven personas and uncovered opportunities for stronger product positioning and pricing strategy.
Using insights from the CEO Report, Data Analysis Plan, and Consumer Insights presentation, I delivered a concise roadmap to help Beats refine its marketing strategy and better meet customer expectations.
This project highlights my ability to translate data into actionable business decisions and create meaningful, insight-driven narratives.

My team and I developed an automated agentic financial workflow using AWS Bedrock that processes raw transaction data into structured insights. The system categorizes expenses, computes KPIs, generates monthly summaries, and performs self-reflection to improve accuracy over time. This project demonstrates how large language models can replicate FinTech processes such as budgeting, transaction classification, and cash-flow reporting while maintaining transparency and auditability.

My team and I created a relational database system designed to help organizations track vulnerabilities, affected systems, remediation teams, and patch progress. The VRT models real enterprise workflows using a normalized schema that links CVE-based vulnerabilities to systems, patches, and team assignments.

Using SQL queries and custom MS Access forms, the system provides visibility into high-risk assets, open critical vulnerabilities, team workload distribution, and maintenance requirements. Real data from the CISA KEV catalog enhances accuracy and realism. This project demonstrates strong skills in database design, cybersecurity concepts, data integrity enforcement, and analytical reporting.

In this project, I built a one-predictor linear regression model in R to analyze the relationship between an independent variable (IV) and dependent variable (DV). After merging two datasets and identifying missing values, I applied multiple imputation using the mice package and bootstrap methods to recover 623 complete observations (page 1). I then fitted an OLS model, confirming a strong and statistically significant linear relationship (p < 2.2e-16) and an adjusted R² of 0.738, indicating that ~78% of variance in the DV is explained by the IV (page 3–4).

In Part B, I performed transformation analysis to improve model fit, identifying log(DV) as the appropriate transformation for correcting nonlinearity (page 2). This raised the adjusted R² from 0.41 to 0.61, and a Lack-of-Fit test confirmed that the transformed model showed no significant LOF (p = 0.137).

This project demonstrates my ability to handle incomplete datasets, apply statistical imputation, perform regression modeling, evaluate diagnostic assumptions, and justify transformation choices using statistical evidence.

In this project, I used R to build and optimize a multiple regression model to analyze gene–environment (GxE) interactions. After diagnosing heteroscedasticity, I applied a Box–Cox transformation to improve model fit and then performed stepwise regression to identify the most significant predictors. The final model included key environmental variables (E1, E2, E4) and a major genetic interaction (G6:G12), achieving an adjusted R² of ~0.5568 and strong statistical significance across predictors. This work demonstrates my ability to perform regression diagnostics, feature selection, and model optimization using R.