How City Maps Are Built in Open World Games (October 2026)

How city maps are built in open world games is a layered data problem, not a modelling problem. A designer plans districts and landmarks, blocks the city out as plain grey geometry, lays a road graph, splits blocks into lots, drops modular building kits onto those lots, dresses them with props, then layers population and traffic on top. The finished world is streamed in chunks so only the nearby cells are ever loaded.

Every open world city you have driven through was assembled in roughly that order, whether the team was three people or three hundred. The order matters more than the tools, because it is what keeps a map from collapsing into a grid of grey boxes with nothing to do.

What Is the Role of a City Map in an Open World Game?

A city map is the playable foundation that holds locations, roads, districts, activities, progression and exploration together. It decides where the player can go, what they find there and how long the trip takes, which means map design sets both the freedom of the sandbox and the size of the production schedule behind it.

The city map is data, not scenery

Newcomers usually picture a map as a big model of buildings. That is the last three percent of the job. Most of what a city map actually is lives in data: spline curves for roads, a list of lots, rules that say which kit piece may sit on which lot, a navigation mesh that pedestrians walk on, and spawn points attached to activity types.

That distinction is why a hand-built city can be a few hundred artists’ assets but millions of placed instances. The art is reusable; the placement is specific.

Layer	What it holds	Who owns it	Built when
Terrain and water	Heightmap, rivers, coastline, beaches	Environment art and world design	First, before anything sits on it
Road network	Road splines, junctions, lanes, bridges	Level design	Early, once terrain settles
Zoning and lots	District rules, block subdivision, building footprints	City design	Alongside the road plan
Modular buildings	Kit pieces assembled into façades and interiors	Environment art	After the blockout is proven
Set dressing	Props, foliage, signage, clutter, decals	Level art	Once the shapes are right
Gameplay	POIs, missions, shops, garages, checkpoints	Design and gameplay	Once a district is walkable
Simulation	Traffic, pedestrians, patrols, economy	Engineering and AI	Last, on top of a finished shape

How City Maps Are Built in Open World Games

The build runs in a fixed order: concept and scale first, then a grey blockout, then roads, zoning and lots, then modular buildings, then props and landmarks, then navigation and simulation, then streaming and optimization, then playtesting. Teams jump backward constantly, but almost never forward past a stage that has not been playtested yet.

The nine-step pipeline

1. Pick a real-world scale and decide the map’s playable area in square kilometres.
2. Write the district plan: how many districts, what each one is for, and where they sit.
3. Choose landmark anchors that players will use for navigation long after names are forgotten.
4. Block out the city as untextured grey massing to test scale, sightlines and travel time.
5. Lay the road graph, then subdivide each block into lots and assign zoning rules.
6. Place modular kit pieces onto lots, then fill facades and interiors with kitbashed variants.
7. Dress streets with props, foliage, signage and lighting to break repetition.
8. Bake the navigation mesh and set traffic, pedestrian and spawn rules on top of it.
9. Stream the world in chunks, then playtest traversal, performance and unfinished areas.

Steps one through three are cheap and wrong-able. Steps six and seven are where the money goes. If the grey massing in step four feels boring, no amount of prop work rescues it, which is why good teams refuse to move past blockout until someone has driven the layout.

The design map is not the playable map

Designers draw a clean diagram with tidy district boundaries and short labels. The playable world is that diagram at ground level, where the label downtown is a two-hour drive from the starting point and the river you drew as one neat curve turns out to have four bridges on it. Both maps have to be true at once, and they never match perfectly.

Hand-built districts versus procedural ones

Most shipped cities are a mix, and the argument between the two approaches is really about where the repetition lives.

Factor	Hand-built district	Procedurally generated district
Control	Total; the designer decides every corner	Rules decide, designer sets the rules
Production time	Slow per district, scales with people	Fast to fill kilometres of land
Uniqueness	Every block can be different	Repetition is the failure mode players notice
Iteration cost	Expensive to change after art lands	Cheap to re-roll, expensive to art-direct
Narrative fit	Easy to stage a story in a known place	Harder to guarantee a good set piece

The complaint you see most often from players, and from developers reading their own playtests, is that procedural cities look samey. Randomising the order of the same building library produces variety in arrangement but not in shape, and players read that as fake within minutes.

How Do Developers Plan the City’s Scale and Layout?

Scale comes first because everything after it inherits the choice. Pick a real-world reference, decide how much of the city is playable, and check that the numbers hold at walking speed and at driving speed.

Pick a real-world scale first

Teams usually anchor the map to something honest, such as a real district compressed, or a real city with the outer ring trimmed. Compressing distances is a legitimate trick, but it has a limit: when two narrative locations that should be an hour apart sit four minutes apart, immersion breaks. That compression problem is the single most common design complaint aimed at modern open worlds.

Test the layout by timing it. A district that reads as fifteen minutes of driving on the map can be four minutes in an average car, and the fix is usually distance or a deliberate traffic jam, not a bigger number on the map screen.

Use landmarks as wayfinding

Players navigate by silhouette. A tower, a stadium, a refinery stack or a suspension bridge visible from three districts away does more work than a full street-sign system, and it costs one asset instead of ten thousand.

Place anchors so that each district has one visible from its neighbour, then check the sightlines in the blockout. If two anchors overlap from the main road, one of them is wasted art budget.

How Do Terrain, Water and Coastlines Shape the Map?

Terrain decides the map before anyone draws a road. Heightmaps built in a landscape tool define where the city can and cannot go, and the coastline is the cheapest large-scale boundary a designer can have.

Flat ground suits grids, which is why most dense downtowns are drawn on plain terrain with the slopes pushed out to the edges. Hills and rivers do the opposite: they break the grid, force bridges, and give districts a natural identity.

Water also acts as a hard barrier. A harbour, a lake or a cliff line saves the team from building a wall and gives the player a clear mental edge for the map. Coastlines with beaches are more expensive because sand, tide lines and access routes all need art and collision.

How Are Roads, Intersections and Traffic Networks Created?

Roads are curves first and geometry second. A road spline carries lane count, width, surface type and speed limit, and the mesh is generated along it. Intersections are authored objects that stitch splines together and define who has right of way.

Laying the road graph

Start with arterials that connect the landmarks, then hang local streets off them, then add the awkward connectors: on-ramps, tunnels, bridges over water and rail crossings. Every one of those is a collision and pathfinding problem later, so they get reviewed early rather than discovered in play.

Lane layouts matter more than they look. Traffic simulation reads lane direction and connection data, so a mis-signed lane produces vehicles that turn against traffic or sit in a junction forever.

Keeping traffic from gridlocking

Traffic density is tuned against the road graph, not by taste. Too many vehicles on a two-lane road turns a main street into a parking lot, and players blame the map even though the map is fine.

Three failures show up in almost every playtest: routes that dead-end at a barrier, intersections where two busy roads meet without a light, and shortcuts that let a player cut through a district faster than the design intends. The last one is fixed with geometry, not rules, and it is cheaper to fix during blockout.

How Do Buildings and Districts Get Placed?

Buildings start as footprints. The blockout team draws boxes at the right scale and height so designers can judge massing, then the art team replaces them with modular kits that snap to those footprints.

Modular kits and kitbashing

A modular kit is a small set of pieces designed to connect: a bay of facade with a fixed window spacing, a corner piece, a ground-floor shopfront, a rooftop unit. Because every piece shares dimensions, one kit covers hundreds of buildings, which is how a city of thousands of structures stays affordable.

Kitbashing is the art of combining pieces across kits and kitsbashed sets so no two facades read identically. A cornice from an industrial kit on a 1920s office block is cheap, invisible to most players, and the reason a street does not feel stamped out.

Interiors are separate worlds

Anything the player enters gets built at a different standard. Interiors need doors that work, rooms that match the exterior footprint, and lighting that matches the time of day outside. Many teams build exteriors years ahead of interiors, then discover that a promised hideout has no back wall.

Zoning ties it together: rules assign each lot a type, and each type decides which kits, heights and prop sets are allowed. Changing a district from commercial to industrial then becomes a rule edit rather than a rebuild.

How Do NPCs, Vehicles and Activities Fill the City?

The finished geometry does nothing on its own. A navigation mesh tells pedestrians and cops where they can walk, spawn rules decide when an agent exists, and activity anchors decide what each corner of the map is for.

Agents are spawned when the player is near and despawned when they are not, so a whole city of pedestrians costs only what the streaming system has loaded. Missions, shops, garages, hideouts tucked beside them and checkpoints are placed deliberately, usually a drive apart, so travelling between them has something to look at.

Population is a dial, not a constant. Too dense and the city feels alive but unplayable; too sparse and streets read as empty at night. Most teams tie density to time of day and district type, which is also how they make the same blockout feel different between a noon drive and a midnight one.

How Do Developers Test and Optimize an Open World Map?

Testing an open world is mostly about two questions: can the player get anywhere, and does the machine survive the attempt.

Traversal playtests are done on foot first, then by vehicle, then by every vehicle class the game has. Designers log where a player got stuck, where they got lost and where they simply turned around because the map had nothing to show them.

Technical testing covers frame rate in the heaviest district, memory use after long travel, loading times when crossing a chunk boundary, draw calls with a full street in view, and traffic density on the worst-case road. If a streamed world stutters while a helicopter flies over downtown, the stutter is the bug, not the city.

Unfinished-area checks are unglamorous and essential: search the map for untextured surfaces, missing collision, placeholder props and doors that open onto nothing. Shaders can catch some of this automatically, and artists can catch the rest faster than any tool.

Real developers report that performance, not content, is what takes longest to fix in an open world. That matches what most teams say out loud: the world is finished when it holds frame rate, not when the last building is textured.

How Does a Small Mod Map Differ from a Full Game City?

A mod map inherits everything the base game already built: the engine, the streaming rules, the traffic system and the file budget. That constraint is the whole difference. A shipped city is designed and tested over years, and a custom map has to fit inside structures that were not designed for it.

Scope is the first casualty. A district that looks modest in a top-down screenshot can be three or four times the work once entrances, interiors, collision and traffic routes are handled. Most mod maps that fail, fail at the interiors and the entrances rather than at the skyline.

Tools modders actually use

Mapping editors built for existing games handle the geometry, and the underlying engine handles lighting, traffic and streaming through those editing tools. Real-world data is a common starting point: roads, building footprints and elevation from OpenStreetMap can be imported, cleaned up by hand, and used as a reference rather than shipped as-is, since the raw data has no game-ready geometry.

Collision is the other recurring problem. Anything a player can stand on needs collision that matches what they see, and getting that right without engine access is where most custom maps run into trouble. Compatibility matters too, because a map built for one game version can break on the next patch if the mod tools changed.

Our advice for anyone starting: build one block, make it fully enterable, and get that single block running at frame rate before planning a district. It is much less discouraging than finishing a skyline and then discovering the interiors will not load.

Frequently Asked Questions

Conclusion

Open world cities are built in a fixed order: scale and districts, grey blockout, road graph, zoning and lots, modular buildings, set dressing, navigation and simulation, streaming, then playtesting. Teams that skip ahead to art spend their budget fixing shapes, and teams that skip the blockout end up designing a city nobody could walk.

If you want to try this yourself, do one thing first. Fix the map’s real-world scale, choose three landmarks players can see from a distance, block the layout out in plain grey boxes and drive it end to end. Everything after that step, roads, kits, props, traffic, is a refinement of a layout you have already proven works. And if you play open worlds more than you build them, knowing how city maps are built in open world games is a handy way to read a map like a designer. When a district feels believable, ask which landmark is guiding you and which kit the facades are drawn from. When a city feels samey, look for repeated pieces and procedural blocks. That is all it is: layers, in an order, chosen on purpose. This guide was refreshed for 2026.