Galaxy Stats processes tournament data from galaxy.fun to generate comprehensive meta analysis reports. We analyze winning decklists from gauntlet finishers to uncover:
What is Galaxy Stats?
- Card popularity across the meta
- Natural card archetypes through clustering
- Captain-specific synergies and signatures
- Optimal deck building recommendations
How the Analysis Works
Data Collection
We parse public tournament results from galaxy.fun, extracting usernames, captains, and winning deck compositions. Only aggregate statistics are analyzed — individual player data is used solely for decklist and captain analysis.
Card Archetypes
Archetypes are manually curated based on co-occurrence patterns — cards that frequently appear together in winning decks are grouped into strategic archetypes (Treasures, Candy, Mage, Pirates, Animals, Toys, Boons, Fae). While hierarchical clustering helps identify natural groupings, the final classifications are human-defined to reflect meaningful playstyle patterns.
Captain Analysis — Lift
Lift measures which cards are over-represented with specific captains. It compares how often a card appears with a captain versus how often it appears overall. A lift of 2.0× means the card appears twice as often with this captain than expected. This reveals each captain's unique "signature cards."
Best 12 Analysis
For each captain, we identify the 12 most commonly played cards across all winning decks piloted by that captain. This represents data-driven optimal picks for deck building.
What You'll Find in Reports
Top Cards: Overall popularity across all winning decks — these are the cards players add voluntarily on top of the default set.
Card Archetypes: Natural groupings based on co-occurrence patterns, showing which cards are played together.
Captain Signatures: Lift-based associations revealing unique synergies between captains and specific cards.
Best Picks: Frequency-based recommendations per captain for optimal deck building.
Technical Stack
BeautifulSoup4 — HTML parsing
NumPy — Matrix operations
SciPy — Hierarchical clustering
Python — ETL pipeline
GitHub Pages — Static hosting
Open Source
Galaxy Stats is open source. The ETL pipeline, analysis scripts, and report generation code are available on GitHub.