I build research-grade data systems for messy administrative datasets and use them to study higher-education finance, student aid, college pricing, and macroeconomic questions.
M.A. in Quantitative Business Economics and B.S.B.A. in Economics with a theatre minor, University of Nevada, Las Vegas. The work sits across economics, data engineering, empirical research, and technical communication: I write the pipeline, run the regression, and ship the platform.
Open to Economist, Research Analyst, Data Analyst, and Financial Analyst roles. Remote, hybrid, or relocation.
The 2004–2024 consolidated IPEDS analysis file is 147,782 rows × 2,676 columns. Labeled release full-panel-labels-v1 is 141,711 rows × 2,721 columns. Builds run a 39-check acceptance audit.
Open the curtain — program notes
I immigrated to the United States from the Philippines at 14.
Founder and sole employee of Mark Jayson Nation LLC (Las Vegas, formed October 2026). The house ships Macro by Mark, an economics learning and U.S. macro research platform; getriff, an iOS workout app in TestFlight; and a Macro by Mark island-tycoon game now in development.
I edit UNLV's Undergraduate Economics Working Paper Series, which I helped build.
Research focus: higher-education pricing, financial-aid packaging, federal cost-of-attendance policy, panel econometrics, causal inference, dynamic specifications, and large-scale data validation.
- Published "Revisiting Bennett's Hypothesis" in Spectra Undergraduate Research Journal (2025)
- Presented at WEAI's 100th Annual Conference
- Authored 4 working papers on higher-ed pricing, financial aid packaging, and federal cost-of-attendance policy
- Helped build, and now edit, the UNLV Undergraduate Economics Working Paper Series
At UNLV I supported instruction in economics, statistics, econometrics, and senior capstone research: lectures and review sessions for 200+ students a semester, course materials, high-volume office hours, and 10 undergraduate researchers mentored through end-to-end empirical projects.
Indicators, release calendars, dashboards, forecasting, and reference workflows. Integrates FRED, BEA, BLS, Census, and World Bank, with Marco, an AI macro assistant, and a status page built to fail loud, never lie green.
iOS workout app, in TestFlight.
A Macro by Mark island-tycoon game is in development.
Reproducible Python and SQL pipeline from NCES/IPEDS Access databases to unbalanced UNITID-by-year panels. The 2004–2024 consolidated analysis file is 147,782 × 2,676. Labeled release full-panel-labels-v1 is 141,711 × 2,721.
Downloads, harmonizes, panelizes, QA-checks, and exports 25+ years of Federal Student Aid Title IV volume reports into institution-by-award-year panels for grants, campus-based aid, and loans.
Links those FSA volume reports to IPEDS institutions, aligns aid periods, and exports combined research panels.
Python and DuckDB pipeline that downloads, harmonizes, QC-validates, and panelizes NCES IPEDS into research-ready institution-year datasets.
Public replication code for research on cost-of-attendance headroom and institutional grant aid, built from the IPEDS panel. Baseline sample: four-year Title IV public and private nonprofit institutions, 2009–2023.
Personal site, Astro on Cloudflare Pages. Live at markjayson.com.
Creative portfolio, Astro on Cloudflare Pages. Live at creative.markjayson.com.
Open the curtain — how a panel gets made
IPEDSDB_Panel extracts annual IPEDS Access databases with mdb-tools, harmonizes them in DuckDB, and writes a research-ready UNITID-by-year panel. The extended 2004–2024 consolidated analysis file is 147,782 rows × 2,676 columns. The labeled release full-panel-labels-v1 is 141,711 rows × 2,721 columns. A 39-check acceptance audit gates the build.
FSAVolumeReports_Panel keeps the Federal Student Aid side at the OPEID-by-award-year grain: acquisition, identity normalization, program harmonization, and suppression handling for grants, campus-based aid, and loans, across 25+ years of Title IV volume.
IPEDS-FSA_Panel consumes both outputs, resolves annual UNITID links, and exports the combined research panels. It does not rebuild either upstream panelizer.
IPEDS_Paneling is the separate Python and DuckDB lane for downloading, harmonizing, and QC-validating NCES IPEDS into institution-year datasets.
COA_FINAID_SUBS is the public replication layer for cost-of-attendance headroom and institutional grant aid. It starts from the clean IPEDS panel, builds the 2009–2023 four-year public and private nonprofit sample, and estimates institution-level fixed effects. Those estimates are associations, not student-level packaging results.
Open the curtain — house lights
Macro by Mark's status page probes the app's own endpoints and folds in vendor health. Strict-mode checks return a real failure when configuration is broken, instead of a cheerful 200. A sentinel validates the chat config without calling the model. Vendor probes time out, cache, and degrade to "unknown," so a third-party outage cannot take the page down. The monitoring vendor is an environment variable (betterstack, instatus, statuspage, or none), and indicator groups roll up worst-of-fold.
| Area | Tools |
|---|---|
| Data engineering | Python · pandas · DuckDB · SQL · PyArrow · mdb-tools |
| Econometrics | Stata · R · Excel |
| Front end | Next.js · TypeScript · React · Tailwind · Zustand |
| Infra / services | Supabase · Vercel · Trigger.dev · Upstash · Stripe |
| Data sources | NCES/IPEDS · Federal Student Aid · FRED · BEA · BLS · Census · World Bank |
Portfolio sites are Astro on Cloudflare Pages.
The grid is redrawn from the public contribution calendar by a scheduled Action. The count in the card is GitHub's, and it updates when the calendar does.
Open to Economist, Research Analyst, Data Analyst, and Financial Analyst roles. Remote, hybrid, or relocation.
