A vetted catalog of statewide metrics
An inventory of statewide indicators, each traced to its source and linked to a data release calendar, so every team cites the same numbers.
- Data governance
- Excel
- R
Projects
Things I've built and found, from statewide data infrastructure to graduate research and personal tools. I build with AI-assisted development: I set the requirements, make the data decisions, and test and debug what ships.
Policy researchData analysisData infrastructure
Texas 2036
Data infrastructure and research support at a nonpartisan policy organization. These tools are internal, so I describe them only in general terms.
Self-service data tools
Tools that let policy staff pull vetted numbers on their own, with a review step before anything is published.
- Python
- React
- APIs
A de-identification tool
A tested tool, first in R and then in Python, for preparing files with personal information for public release, documented for staff who don't code.
- R
- Python
- Automated testing
- Disclosure control
Texas Appleseed
Consumer protection research on data privacy, debt collection, and court costs. Two reports from this work are forthcoming.
Texas Appleseed, 2026
A field study of Texans' privacy rights
Report forthcoming- Qualtrics
- JavaScript
- Google Apps Script
- R
- Text analysis
- Survey design
- The question
- Texas privacy law gives people the right to see and control the data companies hold about them. Does that right work when someone actually tries to use it?
- What I built
- A text analysis of the state's data broker registry identified brokers that process precise location data. I recruited and trained 16 participants to file requests with a sample of brokers and controllers. They documented each step through a Qualtrics instrument I built with custom JavaScript and an Apps Script backend that assigned anonymous IDs. An R pipeline turned about 160 reports into findings, without ever collecting the contents of anyone's data.
- The result
- Nearly half of reports found privacy-rights information hard or impossible to locate. Only 31% of concluded requests produced any data, and no report named who had bought it. I presented these preliminary findings and three recommendations in invited testimony to the Texas House Committee on Delivery of Government Efficiency.
Texas Appleseed, 2026
Debt-collection lawsuits in all 254 counties
Report forthcoming- Python
- Selenium
- R
- sf
- Census API
- Choropleth mapping
- What I built
- A Python and Selenium collector for the state court reporting portal, which exports one report at a time through a legacy browser mode, with resumable runs and a gap audit that refills missing files. An R pipeline then joins seven years of justice court records with Census population, Federal Reserve household debt-to-income bands, and Urban Institute data on debt in collections.
- The analysis
- Lawsuits per 1,000 residents, an effective default-judgment rate that excludes dropped and dismissed cases, and an outlier framework that tracks which counties stay at the top year after year. I verified every figure in the report line by line against the source data before it went to review.
- The result
- A statewide report for legislators on debt lawsuits and default judgments, forthcoming.
Texas Appleseed, 2026
Precinct maps for community events
- R
- sf
- tigris
- Cartography
- HUD crosswalks
- What I built
- Print-ready poster maps of debt lawsuits by justice court precinct, showing filings, default judgments, default rates, and lawsuits per 1,000 residents, for community events covering two multi-county regions: Dallas, Tarrant, Collin, Denton, and Rockwall counties, and Harris, Fort Bend, Montgomery, and Waller counties. I also built ZIP-level maps of high-cost lender locations from state licensing data.
- Why it matters
- Residents and local partners could see where lawsuits concentrate in their own neighborhoods.
Texas Appleseed, 2026
Harris County fines and fees catalog
- Statutory research
- Data verification
- Excel
- What I built
- A catalog of all 101 active fines and fees across Harris County civil courts, tagged by who sets each one, where it falls in a case, and whether it can be waived for people who can't afford it. An AI-assisted sweep came first, then line-by-line verification against statute.
Employer survey on earned wage access
A Qualtrics survey of Dallas-Fort Worth employers on payroll flexibility and earned wage access.
- Qualtrics
- Survey analysis
Surveillance pricing legislation tracker
A comparison of enacted and pending surveillance pricing bills in more than 20 states, focused on how each defines the practice.
- Legislative research
Earlier research
Infrastructure, transportation, education, and local government, from 2023 to 2026.
Changing city limits
First author of a working paper on how annexation reshapes demographics and political participation in Texas. I built a 600,000+ observation Census block dataset with a population-scaled buffer for extraterritorial jurisdictions, joined to turnout records. The study centers on the city of Galveston and on Cameron County, and I conducted its field interviews in Brownsville and Harlingen.
- R
- ArcGIS Pro
- Census Data API
- Field interviews
Water and wastewater vulnerabilities
Regulatory research on vulnerabilities in California's water and wastewater utilities, which informed a national-security war game for senior federal officials.
- Regulatory review
- Scenario design
Equitable transportation decision-making
Co-authored national guidance for the National Cooperative Highway Research Program, building the equity-evaluation framework and reviewing 32 state and regional transportation agencies.
- Qualitative analysis
- Policy review
Every Texas public-school student
Analysis of statewide student records from campus to state level, a return-on-investment method for the five-county Central Texas education and workforce pipeline, and briefs built for each legislator's district.
- R
- Python
- SAS
- SQL
Computer science course access
A school-level catalog of computer science course availability across several states, aggregated to district and state levels for a national report.
- Sampling design
- Data collection
Graduate research
Projects from my M.A. in Government at UT Austin, from neural networks to survey experiments.
Economic connectedness and voter turnout
Tested whether counties with more cross-class friendships vote at higher rates, joining 2020 precinct returns aggregated to counties with social capital measures and Census controls nationwide.
- R
- Regression
- ACS
- Social Capital Atlas
Reading construction from space
Geocoded Phoenix building permits, pulled satellite imagery for each site, and trained a convolutional neural network to classify whether construction was visible.
- R
- Keras
- TensorFlow
- Google Static Maps API
Gentrification and support for transit
A survey experiment with 462 completed responses testing how framing new transit construction affects support, comparing ZIP codes at risk of gentrification with others. At-risk ZIP codes came from Urban Displacement Project typologies for Atlanta, Chicago, Denver, Los Angeles, San Diego, Seattle and Tacoma, and the San Francisco Bay Area and Sacramento.
- Qualtrics
- Experimental design
- R
Redistricting commissions and fair maps
A 50-state analysis of whether the type of body that draws congressional maps predicts how proportional the results are in the 2020 cycle.
- R
- Regression
- Census data
Regional development plans
An economic development plan for the Brownsville metro and a site selection memo for the St. Louis region, both built on workforce, logistics, cost, and quality-of-life data.
- ACS
- BLS
- BEA
- Location quotients
Redistricting simulations and text mining
Analyzed income and education across simulated redistricting plans for Massachusetts and New York, and built a keyword search across a decade of state news archives.
- R
- redist
- Text processing
Personal builds
Tools I build on my own time. The data decisions behind them are the same ones I make at work.
Texas data visualizer
Open the visualizer. An interactive map of Texas from Census blocks to the whole state, with a 3D view that raises each area by its value. It maps more than 45 measures across three Census periods from 2010 to 2024, harmonized to the same boundaries, and adds transit lines, power transmission, and water service areas. Guided tours step through each region’s metros in 3D, every Texas rail system, and story topics like rent burden and broadband, with fact cards that compare each place with Texas and summarize its spatial pattern (core versus edge, Getis-Ord hot spots, Moran’s I). The Analyze tab runs indices, scatterplots, robust regression, LISA clusters, and rankings on every area in the state, with plain-language interpretations, and exports to CSV, Excel, Stata, SPSS, SAS transport, GeoJSON, and shapefile with a codebook and citation. Statistics were checked against PySAL and statsmodels. R prepares the data, tippecanoe builds vector tiles, Cloudflare R2 serves them, and MapLibre draws them in the browser with no application server.
- R
- tippecanoe
- PMTiles
- MapLibre
- JavaScript
- Spatial statistics
Austin boundary atlas
A Python pipeline that assembles every layer of government over the Austin metro, from city limits and ETJs to school, water, legislative, and commissioner districts, as the base for showing who governs any address.
- Python
- Census TIGER
- Open data portals
Tureen
A shared household recipe app. I designed its relational schema and the household access rules the database enforces, so members see only what is shared with them, and an import flow that normalizes messy recipes into one schema and flags low-confidence fields for human review before saving.
- React
- Vite
- Supabase (Postgres)
- Row-level security
- Serverless functions
Personal finance dashboard
A desktop app for macOS that keeps a Google Sheet as the single source of truth and turns it into spending, income, investment, and scenario views. It has its own test suite, and the same pattern as a policy dashboard: one authoritative source and a reproducible transformation.
- Python
- Google Sheets API
- py2app
- Automated testing