Technical Skills
Languages & tools: Python (pandas, scikit-learn, Plotly), R (glmnet, time series), SQL, Excel, Git / GitHub
Methods: time-series forecasting, multiple regression, LASSO and model selection, k-means clustering, bootstrap and cross-validation, survival analysis (Kaplan–Meier), anomaly detection, data-quality auditing
Independent Data Analysis Projects | 2026
When Traffic Breaks DownCaltrans PeMSPython
- Scripted restart-safe collection of 120 daily Caltrans PeMS files (~2.2 GB); audited detector coverage and restricted analysis to fully observed intervals.
- Compared three breakdown definitions that found 1,457, 10 and 7 events in the same four months; tested results with matched controls, leave-one-out checks and Kaplan–Meier recovery analysis.
- Located a recurring bottleneck: 172 persistent queue runs between two adjacent stations vs. 36 for the next-highest pair.
Ames Housing RegressionR / Python replicationR + Python
- Replicated a full regression workflow in R and Python on identical train/test splits; selections were perfectly nested and test RMSE agreed within $53 on ~$21,000.
- Reduced a 221-column design matrix with LASSO and BIC selection; compared candidate models with a 10,000-resample paired bootstrap.
Electricity Demand ForecastingEIA-930, CaliforniaR
- Compared seasonal-naïve, Fourier + ARIMA, 50/50 ensemble and published EIA forecasts on 11 years of hourly demand (97,344 observations).
- Ensemble cut MAE 11% vs. the seasonal-naïve baseline across 8,544 forecasts; traced 84% of the EIA forecast’s net underforecast to nine daytime hours.
Customer Lifecycle SegmentationUCI Online Retail IIPython
- Cleaned ~1.07 million transactions and reconciled 1,210 same-day purchase-cancellation pairs before building account-level features.
- Segmented 5,823 accounts with k-means, validated with silhouette, Davies–Bouldin and bootstrap ARI (0.981); 18% of accounts produced 72.8% of net revenue.
SQL Business MetricsAdventureWorksSQL + Python
- Queried three years of AdventureWorks sales in SQL with Python / Plotly reporting; online orders were 87.9% of orders but only 26.7% of merchandise revenue.
Professional Experience
NASA Armstrong Flight Research Center / SOFIACal Poly STAR ProgramSummer 2013
- Started with no assigned project; identified and scoped a Telescope Assembly sensor study by asking SOFIA personnel what needed analysis.
- Wrote all analysis code in R and compared system behavior across 14 flights, extending earlier per-flight analysis to separate brief spikes from sustained Power Bus anomalies.
- Determined a threshold for warning software to alert operators to power-load problems; applied the same method to gyroscopic drift.
- First author of the resulting research poster; wrote the initial draft and all figures, with review from NASA, DSI and USRA collaborators.
iLEAD SCVi, Santa ClaritaMath Teacher, Grades 8–11Aug–Dec 2025
Endeavour Middle School, Lancaster7th Grade Math TeacherJun 2024–Jun 2025
Antelope Valley High School, LancasterHigh School Math TeacherAug 2023–Jun 2024
Orange County Dept. of Education, Juvenile HallSubstitute Teacher2022–Aug 2023
Scoot Education, Los AngelesSubstitute / SpEd Paraeducator2022–2023
WyzAntIndependent Tutor, Mathematics & Statistics2009–Present
Education
California State University, FullertonB.A., Mathematics2013