Ernesto Bayon · Data Science & Analytics

I turn messy data into
decisions worth trusting.

MSBA candidate at Pepperdine building applied data science projects — causal inference, forecasting, regression analysis, and full-stack analytics products — end to end, from raw data to deployed app.

About

I'm Ernesto Bayon, an MSBA student at Pepperdine focused on applied data science — the kind that ships. My work spans causal inference and experimentation (pre-registered A/B tests, DoWhy backdoor adjustment, Double ML, propensity score matching, difference-in-differences), statistical modeling (regression, SPSS pipelines, forecasting with ARIMA/ETS), and full-stack product engineering (Next.js, Supabase, the Claude API) to turn analysis into tools people actually use.

Below is a working sample of that range: a production AI task manager that doubles as a live research platform, an economic indicators dashboard built on Mexican public APIs, a retail category regression study, and a handful of product/UX and applied-analytics prototypes.

Projects

Pathfinder

Flagship

An AI task manager for data scientists with a PM-agent chat — and simultaneously a live, pre-registered causal inference experiment testing whether AI-assigned task priority beats a due-date baseline (t-test, DoWhy, Double ML/EconML, PSM, DiD).

Next.jsTypeScriptSupabase Claude APICausal Inference

FinData MX

Dashboard

An economic indicators dashboard built on Mexico's public data infrastructure — Banxico SIE, INEGI BIE, and CNBV APIs — with ARIMA/ETS forecasting on top of the cleaned time series.

PythonTime SeriesPublic APIs

Snack King Analytics

Regression

A salty-snacks retail category case study: cleaning syndicated sales data, exporting labeled SPSS datasets, and running a regression analysis to identify the real drivers of category dollar sales.

PythonSPSSpandas Regression

Rent Value Analyzer

Analytics

A rental value analysis tool for California metros, blending Census and FRED economic data with amenity-weighted scoring (schools, transit, parks, hospitals) to estimate neighborhood rental value.

PythonStreamlitCensus API FRED API

Uber Eats Analytics

Analytics

An exploratory and applied analysis of Uber Eats order data, built as an interactive Streamlit app with cleaned data, business recommendations, and a documented deployment path.

PythonStreamlitpandas

NestWorth Prototype

Product / UX

An interactive UX prototype for a personal budgeting application — dashboards, multi-line cash flow visualizations, and smart filters designed to make personal finance legible at a glance.

PythonProduct DesignData Viz

NorthPeak Excel Profiler

Reporting

A strategic-intelligence report generated from Excel-based business data profiling — a self-contained, shareable snapshot showcasing how a raw spreadsheet becomes a presentation-ready analysis.

ExcelBusiness Intelligence

Stack

Python TypeScript / Next.js SQL / Supabase pandas / numpy SPSS Causal Inference (DoWhy, EconML) Streamlit Claude API Vercel