Senior Data & Business Analyst · Supply Chain & Operations · FinOps
10+ years turning complex data into decisions that move the business forward.
🟢 Actively seeking full-time roles: Senior Data/Business Analyst, Operations/Supply Chain Analyst, or Inventory/Demand Manager
- 📊 10+ years in FinOps, supply chain, and operations analytics at Texas Instruments
- 🎓 MS in Business Analytics candidate at Baruch College (Zicklin), graduating December 2026
- 🧰 I build end-to-end pipelines, from raw API data to production dashboards and statistical models
- 🤝 Always open to connecting, collaborating, or discussing analytics problems
📝 A note on audience: most of the projects below are written for a technical audience, with the methods and code up front. I'm working on companion versions aimed at non-technical stakeholders (shorter, results first, built around the decision someone has to make). The first ones are in the pipeline below.
Business question: Does traffic volume drive infrastructure degradation, and does city response time vary by borough?
Integrates NYC 311 Service Requests and DOT Automated Traffic Volume Counts (2020–2025) into a Kimball star-schema data warehouse. Implements a full ELT pipeline using the Socrata API, Google Cloud Functions, BigQuery, and dbt, then surfaces findings through a Looker Studio dashboard.
SQL · BigQuery · dbt · Looker Studio · Python · Data Warehousing · Kimball · NYC Open Data
CIS 9440 · Data Warehousing and Analytics · Spring 2026
Business question: How well does a third-party signal (Wikipedia pageviews) agree with Netflix's own weekly Top 10 rankings across international markets?
Reconciles first-party (Netflix Top 10) and third-party (Wikipedia pageviews) viewership signals across six international markets over 12 weeks, covering 378 titles and 641 title-market pairs. Maps titles through Wikidata, loads into BigQuery, and measures market-level agreement with week-level bootstrap confidence intervals. Closes with six measurement recommendations for a media analytics team. (Code)
Python · BigQuery · SQL · Tableau · Quarto · Wikidata API · Bootstrap Inference · Data Reconciliation
Business question: When a large facility is proposed near homes, what does living next to it cost, and who ends up living closest?
Starts from a live dispute: a developer offered 4,500 Pennsylvania households $10,000 each to accept a 1,300-acre data center. Uses 2020 Census tract data, ACS median household income, and NYC's PLUTO database to map industrial land across all five boroughs and measure distance from every census tract to the nearest industrial lot. Median household income rises with distance (about $64K within a quarter mile, about $90K beyond a mile), a sorting pattern rather than a measured cost. Pairs that with a published Dutch study of 70,684 home sales (de Vor & de Groot, 2011) that found roughly 15% lower prices within 250 meters, fading by about 1.4 miles, to show why a flat payment misses on both distance and who bears the loss. Includes a data correction note: the first version hit an API row limit and overstated coverage gaps, and the published numbers reflect the corrected pull. (Code)
R · sf · tidycensus · ggplot2 · QGIS · Socrata API · Spatial Analysis · Distance-Band Analysis · Wilcoxon Rank-Sum · NYC Open Data
Business question: Do lower-income neighborhoods wait longer for city services, and if so, is that driven by differential treatment or by what they report?
Analyzes 12.6 million closed NYC 311 service requests (2022–2025) spatially joined to census-tract income data. The raw gap is real (43% longer for the lowest-income quintile), but OLS, bootstrap confidence intervals, and quantile regression show the disparity is driven by complaint type composition, not differential treatment. Part of a four-person capstone (joint summary report).
R · sf · tidycensus · leaflet · quantreg · Spatial Analysis · Regression · NYC Open Data
STA 9750 · Software Tools for Data Analysis · Spring 2026
Business question: What patterns in health code violations and closure rates should restaurant operators and regulators prioritize?
Analyzes over 160,000 NYC Department of Health restaurant inspections (2021–2024), uncovering borough-level grade differences validated with ANOVA, closure rate trends, and the most common violations by cuisine category.
Python · Quarto · Data Analysis · NYC Open Data · Public Health
CIS 9650 · Programming Tools for Analytics · Fall 2025
Business question: How have demographic trends at U.S. colleges shifted following the Supreme Court's affirmative action ruling?
Data-driven analysis of enrollment trends across U.S. institutions following the Students for Fair Admissions decision, identifying which institution types and demographics show the strongest signal.
R · Data Analysis · Higher Ed Policy
STA 9750 · Software Tools for Data Analysis · Spring 2026
Business question: Which customers are most likely to churn, and what actions should the retention team prioritize?
Binary classification study on telecom customer churn using the Kaggle Telco dataset. Builds and evaluates a logistic regression model with interaction terms, tunes the classification threshold to maximize recall, and translates model coefficients into actionable retention strategies.
Python · Logistic Regression · Classification · Customer Analytics · Telecom
CIS 9660 · Machine Learning for Business Analytics · Spring 2026
Business question: How many gold medals should sponsors expect Team USA to win at the 2028 Los Angeles Olympics?
Builds a four-factor model (US baseline, host nation effect, new sport bonus, GDP per capita) from Summer Olympics data scraped from Olympedia. The host effect is estimated for hosts that were already medal powers, since Team USA's situation looks nothing like a small nation's first home Games. Effects come from bootstrap confidence intervals via the infer package, and a 1,000,000-draw Monte Carlo simulation combines them. The model forecasts 74 golds (95% interval 61 to 88). Back-testing against 10 past hosts adds an average miss of 6.6 golds, which widens the range to 54 to 95. Delivered as a fundraising brief for the LA 2028 Organizing Committee.
R · rvest · tidyverse · infer · ggplot2 · gt · Monte Carlo Simulation · Web Scraping · Bootstrap Inference · World Bank API · Olympedia · Quarto
STA 9750 · Software Tools for Data Analysis · Spring 2026
Business question: Where are Americans moving, what does it mean for congressional representation after 2030, and how should Texas protect its political footprint?
Analyzes American Community Survey migration flow data to identify dominant corridors, quantify Texas's inbound growth, and close with a data-driven advertising strategy for preserving congressional seats.
R · Data Analysis · Demographics · Migration · Political Geography
STA 9750 · Software Tools for Data Analysis · Spring 2026
Business question: How do different demographic groups actually spend their time, and what user profiles should a time-tracking app be built around?
Builds survey-weighted time-use profiles from multiple years of American Time Use Survey microdata, averaging hours per person per day (zeros included) so each activity reflects the whole group, not just the people who did it. Compares groups side by side and turns the differences into product personas for a gamified app. The report was republished after a full correction pass on activity coding, respondent filters, and the averaging method; the original publish date is kept and the republish date is shown in the title block.
R · Survey Weighting · Weighted Quantiles · gt · Quarto · Time Use · Survey Data
STA 9750 · Software Tools for Data Analysis · Spring 2026
Two more projects are in progress and will be added here when they're published:
- ☀️ Solar energy forecasting: a time-series forecasting project on solar energy usage. BUS 9430 · Business Analytics Project Management · Fall 2026
- 🌐 Supply chain resilience: a data visualization project on how global supply chains hold up under disruption. CIS 9655 · Data Visualization · Fall 2026
Both will include a short, results-first version for non-technical readers alongside the technical write-up.
Languages
- Python: Netflix/Wikipedia, NYC Restaurants, Customer Churn, NYC Traffic
- R: Industrial Proximity ($10K), NYC 311, SFFA, Time Use, Migration, Olympics
- SQL: Netflix/Wikipedia, NYC Traffic
BI & Visualization
- Tableau: Netflix/Wikipedia
- Looker Studio: NYC Traffic
- ggplot2 / Quarto: Industrial Proximity ($10K), NYC Restaurants, Olympics, NYC 311, Time Use
Cloud & Data Platforms
- BigQuery: Netflix/Wikipedia, NYC Traffic
- dbt: NYC Traffic
- Google Cloud Functions: NYC Traffic
Statistical & ML Methods
- Logistic Regression: Customer Churn
- Bootstrap Inference: Netflix/Wikipedia, Olympics, NYC 311
- Monte Carlo Simulation: Olympics
- OLS & Quantile Regression: NYC 311
- ANOVA: NYC Restaurants
- Wilcoxon Rank-Sum Test: Industrial Proximity ($10K)
- Survey Weighting: Time Use
- Spatial Analysis (sf, QGIS): Industrial Proximity ($10K), NYC 311
Data Engineering
- Socrata API: Industrial Proximity ($10K), NYC Traffic
- Web Scraping (rvest): Olympics
- Wikidata / Wikipedia API: Netflix/Wikipedia
- World Bank API: Olympics
- Kimball Star Schema: NYC Traffic
Thanks for stopping by! 😊