In This Chapter
The Spark
There's a website that maps musical influence as a network graph. Artists are nodes. Edges connect who influenced whom. Click on a band and you can trace the lineage of an entire genre backward through decades, or forward to see what it spawned. It's a simple concept, but the experience is magnetic. You start with one artist you know, and twenty minutes later you're five hops deep in a web you never expected.
That was the starting point. Not the technology, not the dataset, not even the medium. Just the experience of following influence chains and discovering connections you didn't know existed.
From Music to Games
Music influence graphs work because the data is relatively clean. Artists cite their influences in interviews. Critics trace lineage. The connections are documented, debated, and largely agreed upon.
Games are messier. A game's influences might span mechanics from one title, art direction from another, a level design philosophy from a third. Developers rarely cite their full list of references. Some influences are obvious (every immersive sim traces back to Ultima Underworld); others are invisible without deep knowledge of the craft.
But that messiness is also what makes it interesting. Games don't just influence each other through inspiration. They share studios, engines, publishers, creators. Teams split and reform. Mods become standalone titles. Rivalries push both sides forward. The relationship vocabulary for games is far richer than a simple "influenced by" edge.
The goal became clear early: don't just map influence. Map the full connective tissue of the medium. Studio lineage. Shared engines. Design antithesis (games that exist specifically as a reaction against another). Dev team exodus (when a group leaves one studio and the DNA carries over). Mod origins. Spiritual successors. Rivalries.
Drag any node and watch the simulation respond. Genre clusters emerge from the physics alone; nobody told the layout where RPGs or platformers should go. Tune gravity, repulsion, and link strength to see how the structure changes.
The Third Dimension
The music influence sites use flat, 2D layouts. That's fine for a few hundred nodes. But game history spans thousands of titles across decades, with far more relationship types creating a denser web of connections. In two dimensions, the graph collapses into a hairball.
The first prototype was flat. 2D, like the music sites. It worked with a hundred nodes, but as the dataset grew the graph collapsed into an unreadable tangle. Too many relationship types creating too many crossings. The hairball problem.
The decision to go 3D wasn't about spectacle. It was about survival. A third axis gives the layout algorithm room to separate clusters that would overlap in 2D. Genres can drift into natural neighborhoods. Eras can spread along a temporal axis. The result feels less like a diagram and more like a place you can move through.
2D Layout
Dense graphs collapse into overlapping clusters. Relationship crossings make the structure hard to read. Works for small, sparse networks.
3D Layout
The extra axis gives clusters room to separate naturally. Camera movement lets you approach the data from any angle. Feels explorable, not just readable.
The "star map" metaphor emerged from the 3D layout. When you zoom out, the game nodes look like stars. Clusters of related games form constellations. Dense regions glow. The metaphor suggested the name, the visual language, and eventually the entire design direction.
The Aha Moments
Three things made the project feel worth finishing. The first was watching clusters emerge from the force simulation. Nobody told the layout engine that RPGs should group together, or that 90s shooters form a constellation distinct from modern tactical shooters. The data did that on its own. Seeing structure emerge from raw connections was the first proof that the approach worked.
The second was timeline mode. Scrubbing from the 1970s to the present and watching the galaxy fill in, era by era, felt like watching the medium grow. You could pause on any year and read the state of the industry in the shape of the graph. 1993 looks different from 2003 looks different from 2023. The evolution is visible, not just described.
The third was the screensaver. Let the timeline play on its own with the camera drifting through the galaxy, and the old cover art floating past. There's something about seeing game covers from your childhood appear as points of light in a constellation that makes the whole thing feel worth the effort.
The People Problem
One of the first things that happens when a game developer sees Six Degrees is they search for a game they worked on. Every time. Doesn't matter if the studio shipped AAA blockbusters or a two-person indie. They want to see if they're on the map.
The letdown when they're not there was one of the hardest things about the project. It exposed a fundamental tension in the data model. The first round of relationship types was driven by what IGDB provides: shared studio, shared publisher, shared genre. That's useful for structural analysis, but it misses the thing that actually matters most to the people who make games: the team.
Connecting specific developers to games turns the influence graph into something much larger: an industry map. People move between studios. They carry ideas with them. The DNA of a game isn't just its genre and engine; it's the humans who built it and what they learned at their previous job. Modeling that accurately, without reducing teams to a handful of named leads, is an unsolved problem. It's also the most important one.
Scope and Ambition
The initial prototype was a weekend experiment. Load some games into a force-directed graph, draw some lines, see if it felt interesting. It did. The force simulation naturally grouped games by genre without any manual positioning. RPGs drifted toward RPGs. Shooters clustered with shooters. The emergent structure was enough to prove the concept.
From there, scope grew in layers. IGDB integration for real metadata. Supabase for persistence. Timeline playback so you could watch history unfold. A pathfinder for the "six degrees" party trick. Steam import so you could see your own library in context. Each feature existed because the previous one raised a new question.
Why "Six Degrees"? The name is a nod to the small-world property of networks. In social networks, any two people are connected by roughly six intermediate links. The same property holds in game influence graphs. Pick any two games, no matter how different, and the pathfinder usually connects them in under six hops.
The project took shape during evenings and weekends, built in the gaps between a full-time job at a game studio. That constraint shaped every decision. The architecture had to be low-maintenance. The data pipeline had to be scriptable, not manual. The UI had to work without constant attention. Everything was designed for burst-mode development: intense sessions followed by weeks of quiet.