incoming data manager at Mississippi Engaged
I turn raw data, from ticket prices to live feeds, into something a team can run on. Pipelines in dbt and BigQuery, automation in Python, results in the browser.
For the Pennsylvania Democratic Party's voter-protection program I automated the GOTV packet run end to end, backtested a ballot-cure model with a 3,100-line replay harness, and moved the team's BI layer off Tableau and onto code.
On the side I forecast football attendance from resale ticket prices and publish precinct-level election analysis that people actually read.
research application
A desktop research tool for Jackson Free Clinic that measures access to JTRAN bus service and saves confirmed measurements locally, without retaining entered addresses.
100k+views on primary day
Live wait times at every polling site in New York City during the 2025 mayoral primary, rolled up to Council, Assembly, Senate, Congressional, and Election District.
170,891early votes tracked, 2026 primary
A live early-vote dashboard for Travis County, Texas: daily turnout and partisanship paced against 2024, with a model projection of where the final count lands.
2,229seat leave-one-out RMSE
A Python package and CLI that forecasts Mississippi State home-game attendance from season-relative resale prices. Its first saved live forecast missed by 1,978 and covered the actual crowd in its 80% range.
32special elections, 13 states
A hand-built precinct dataset comparing 2026 special-election results with 2024 presidential margins, generic-ballot polling, and presidential approval at the time of each race.
~530transfers, 2022 to 2024
How well Group of Five players hold up after moving to the Power Five: regression and correlation analysis by position group on PFF data.
dbt projects laid out as staging, intermediate, and marts on BigQuery, with data-quality tests in the pipeline and fuzzy matching where the source systems disagree about who a record is.
Python that replaces manual work: browser automation with Playwright, TOTP-authenticated logins, REST clients packaged as reusable libraries, pytest suites, and jobs scheduled on Google Cloud and GitHub Actions.
Precinct and boundary work in QGIS: matching, consolidating, and geocoding at scale with Census boundary data and batch geocoders, then turning the result into maps a decision can rest on.
Code-based BI in Observable Framework on top of warehouse marts, deployed like software, plus the application layer around it when a dashboard is not enough.
NGP VAN administration and the VAN and MyCampaign APIs, voter file analysis, and turf cutting for field programs, including training the organizers who use it every day.
Mississippi Engaged is a nonpartisan 501(c)(3) that supports nonprofit and community-based organizations across Mississippi with civic engagement coordination, grants, data, and digital organizing. Its partners work on voter engagement, election protection, and building community power. About Mississippi Engaged →
Data Science coursework. CSCI 101, Foundations of Data Science and Engineering, grade A.
B.A., History. Coursework in Statistics, Computer Programming, and Public Policy. Dean's List twice, President's List Spring 2023.
A precinct-level dataset built by hand across 32 special elections in 13 states, comparing each result with the 2024 presidential vote, the generic-ballot average, and presidential approval on election day.
Roughly 530 players who moved from the Group of Five to the Power Five between 2022 and 2024, with regression and correlation analysis by position group on PFF grades.
Reporting on a fabricated poll, borrowed branding, and the accounts behind its release. Originally published by Elections Today in 2020; recovered from the archive and edited for republication.
I join Mississippi Engaged as Data Manager in October 2026, working remotely. Mississippi Engaged is a nonpartisan nonprofit that coordinates and resources civic engagement work across a statewide network of nonprofit and community-based organizations. This site brings together my work in analytics engineering, civic data, and sports analytics.
SQL, dbt, and BigQuery for modeling and warehousing. Python with pandas, pytest, and Playwright for automation. Observable Framework for code-based BI. QGIS and the Census API for geospatial work. GitHub Actions and Google Cloud for scheduling and CI.
Sports, civic data, and operational analytics. I build football attendance forecasts, live dashboards, data pipelines, and geospatial analysis. My case studies cover how the systems work; my articles explore what the data says.
Have a question about my work or want to connect? Email gets the fastest reply.
