Competition Analytics
Explore event, club, and athlete-level performance from structured Brazilian Jiu-Jitsu competition data.
About Smoothcomp Stats
A data project for exploring the competitive landscape of jiu-jitsu.
Smoothcomp Stats is a portfolio-style analytics project built to make Brazilian Jiu-Jitsu competition data easier to explore. The goal is to turn tournament results into structured, searchable, and useful views across events, clubs, and athletes.
About Me
Hi — I’m Dusty.
I work in analytics and data modeling, with a background in actuarial science and a growing focus on data engineering. I like building structured datasets from messy real-world sources and turning them into something measurable, useful, and easier to reason about.
This project combines that technical interest with my interest in Brazilian Jiu-Jitsu. I wanted to understand what the competitive landscape looks like when you zoom out: how athletes progress, how clubs perform, how events compare, and what patterns show up over time.
My broader goal is to make competition data more transparent and less intimidating, especially for newer competitors who are trying to understand tournaments, divisions, and performance trends.
Project Focus
What this site is trying to answer
Data Engineering
Build reliable pipelines for collecting, storing, transforming, and serving analytics-ready data.
Skill Progression
Use metrics, trends, and future rating systems to better understand how competitors improve over time.
Technical Overview
Built as an end-to-end analytics engineering project.
Smoothcomp Stats is not just a static website. It is a full data pipeline: data is collected, stored, transformed, modeled, exported, and served through a frontend explorer. The frontend is intentionally lightweight, but the backend workflow is designed around production-style analytics engineering patterns.
Python-based data collection
Cloud-hosted scraping and batch processing
Amazon S3 for raw and curated data storage
SQL and dbt for analytical modeling
Athena and DuckDB for querying
Astro for the frontend explorer experience
Current Features
What is available now
- Recent event explorer
- Event-level club summaries
- Club summary pages
- Athlete summary pages
- Performance trends by event index
- Repeat-opponent and rivalry analysis
Future Direction
Where the project is heading
- ELO-style competitor ratings
- Opponent-strength adjusted metrics
- Club and athlete rankings
- More visual trend analysis
- Predictive performance modeling
- Better tools for new competitors to understand tournaments
Why this matters
Competition data can be more useful than a list of results.
A single match result tells you who won. A structured dataset can show much more: whether an athlete is improving, whether a club is active across many events, how often competitors win by submission, and how performance changes across experience levels.
The long-term goal is to make this data useful for competitors, coaches, fans, and technical readers who want to understand both the sport and the data engineering behind the platform.
Connect
If you are interested in the technical side of the project, data engineering work, or the analytics behind competitive jiu-jitsu, you can find more of my work on GitHub and LinkedIn.