Portfolio

Josh Carpenter

aka agoodtheory

I'm a Senior Business Analyst with a strong SQL background and a passion for data, programming, and building solutions in the cloud. I'm working toward a career transition into data engineering through hands-on project work and formal AWS training. Outside of work I enjoy building things — from serverless data apps to tinkering with hardware, to whatever idea lodges itself in my head and won't let me rest until I figure it out — and you'll usually find me hiking or fully locked into gaming.


Projects

Data Engineering Graph Theory AWS Python Serverless

Twilight Bark

A full-stack data engineering project that models how a message would travel across the continental US via dog-owning households, county by county. Built on U.S. Census data, a weighted NetworkX graph, Dijkstra pathfinding, and a serverless AWS backend — with an animated Leaflet.js frontend that visualizes the route in real time.

System Architecture Vector DB Serverless AWS Data Pipelines

The Compendium

A private, AI-powered research archive for cataloging and cross-referencing reported paranormal phenomena — Bigfoot sightings, UFO/UAP reports, Missing 411 cases, and more — normalized from public datasets into a unified schema. Entries are embedded into a Pinecone vector index and queried through a Claude-powered reasoning layer that surfaces cross-case patterns: geographic clustering, seasonal timing, behavioral overlap across otherwise unrelated reports. There's no link to it anywhere on this site. Access is triggered by a passphrase typed directly into the AI chat widget in the lower right corner, which silently unlocks a hidden login screen. The architecture is open. The door is locked.

ETL Python Tkinter AWS OpenAI

The Archivist

A standalone desktop ingestion pipeline that feeds The Compendium. Takes CSV, XLSX, or PDF input, fuzzy-matches columns against a configurable field schema, shows a live JSON preview and cost estimate before committing, checks for duplicates against a DynamoDB GSI, and delivers in one pass to DynamoDB, S3, and Pinecone. Packaged as a portable Windows executable.


Skills & Tools

Data

  • SQL (advanced)
  • Python / pandas
  • geopandas
  • NetworkX
  • Excel / Power Query

Cloud & Infra

  • AWS Lambda
  • API Gateway
  • S3 / CloudFront
  • Route 53
  • AWS CLI

Frontend

  • HTML / CSS / JS
  • Leaflet.js
  • Responsive design
  • REST APIs

Tools

  • Git / GitHub
  • VS Code
  • Jupyter
  • Census / HUD APIs

About

The handle agoodtheory has been mine for years — a gaming tag that made its way onto GitHub and eventually into a domain name. It fits: I tend to show up with a theory, test it against reality, and see what survives.

My background is in business analysis, where I've spent years translating data into decisions. I'm now building toward data engineering — combining the analytical foundation I already have with the technical skills to work further upstream in the data pipeline.

When a problem gets interesting enough, I can't let it go. Twilight Bark started as a half-formed idea on a walk and turned into a full-stack data project with a live AWS deployment. That's usually how it goes.

Currently

Senior Business Analyst

Targeting

Data Engineer · Analytics Engineer

Training

AWS (Adrian Cantrill) · Python · Cloud Infrastructure

Outside work

Hiking · Gaming · Building things

josh_bot_v1.0
portfolio assistant
Hi there. I'm Josh's portfolio assistant. Ask me anything about his background, projects, or skills.
Guardians Make Their Own Fate
click anywhere to dismiss
The Compendium Personal Research Archive · v0.1
Active Session
All Entries · 12 records
Research Assistant
Ask across the full archive
// Session initialized. Archive loaded.
The archive is open. What patterns are you chasing?
Full Archive
Browsing all entries
// Loading archive...
Page 1
Map View
Plotting entries with known coordinates
Category
Entry Title
Date of Event
Source
Tags
Added
// Loading entry...