Some of the most important geographies are not places. They are movements between places.
People commute. Goods move through ports. Migrants travel between regions. Aircraft connect cities. Water flows through rivers. Information moves through networks. A conventional location map can show the origins and destinations, but it cannot fully explain the movement connecting them.
Flow mapping represents movement between places using lines, arrows, curves and varying line widths to encode direction, volume, route or intensity.
A flow map turns connection into geometry.
Quick Read: The Flow-Map Mechanism
ORIGIN + DESTINATION + MOVEMENT VALUE → CONNECTION LINE → DIRECTION / WIDTH / ROUTE → VISUAL NETWORK OF FLOWS
1. Every Flow Has an Origin and Destination
A commuting flow begins at one place and ends at another. Trade moves between ports or countries. Migration links sending and receiving regions. Without origin and destination, a line is merely geometry rather than a movement relationship.
2. Direction Matters
Arrows can distinguish A→B from B→A. When flows are asymmetric, removing direction can hide which places are sources, sinks or exchange partners.
3. Width Can Encode Volume
A thicker line can represent more passengers, tonnes of cargo, migrants or messages. As with Proportional Symbol Mapping, the scaling rule must be clear so visual magnitude matches the underlying data honestly.
4. Desire Lines Show Connection, Not Necessarily Route
A straight or curved line between origin and destination may show that movement exists without claiming travellers follow that exact geographic path. Desire lines emphasise relationship rather than route geometry.
5. Routed Flow Maps Follow Networks
When actual roads, railways, rivers or shipping lanes matter, flows can be routed through the network. This connects to Network Distance, which owns movement through constrained route systems.
6. Flow Mapping Is Not Spatial Interaction
Spatial Interaction owns the broader mechanisms that generate movement between places. Flow mapping owns the cartographic representation of that movement.
7. Flow Mapping Is Not Network Distance
Network distance measures separation through a route system. Flow mapping can display movement volumes along or between routes, but it does not itself calculate shortest paths or travel cost.
8. Flow Mapping Is Not Proportional Symbol Mapping
Proportional symbols encode magnitude at locations. Flow maps encode magnitude along relationships between locations. One makes places bigger; the other makes connections stronger.
9. Too Many Lines Create Hairballs
A dense origin–destination matrix can produce thousands of crossing lines. The result may become visually useless even when the data are correct.
10. Aggregation Can Reveal Structure
Flows can be grouped by region, corridor or dominant destination to reduce clutter. But aggregation changes the story: local variation may disappear while larger-scale patterns become clearer.
11. Thresholding Removes Weak Flows
Maps sometimes display only flows above a minimum value. This improves legibility but can hide many small connections whose combined effect matters.
12. Flow Direction Can Be Net or Gross
If 1,000 people move from A to B and 900 move from B to A, the net flow is only 100 even though 1,900 movements occurred. Net maps reveal directional balance; gross maps reveal total exchange. They answer different questions.
13. Time Changes the Flow Map
Morning commuting, weekend travel, seasonal migration and annual trade can produce very different networks. A static flow map can hide temporal variation if the time window is not explicit.
14. Primary Geography: Draw Where Everyone Goes
Ask pupils to draw arrows from home zones to a school. Thicker arrows can represent more pupils. They quickly see that movement patterns can be mapped differently from locations.
15. Secondary Geography: Compare Desire Lines and Actual Routes
Students can draw straight origin–destination lines and compare them with real network paths. The difference reveals where barriers and transport systems reshape movement.
16. Advanced Geography: Origin–Destination Matrices
Flow data are often stored as matrices where rows represent origins, columns represent destinations and cells contain movement volume. Mapping converts this relational table into spatial structure.
17. Singapore Example: Commuting
Flows between residential towns and employment centres can reveal directional commuting corridors. A morning map may show strong inbound movement toward major job centres, while an evening map reverses many arrows.
18. Singapore Example: Port and Airport Networks
Shipping and aviation flows connect Singapore to regional and global nodes. Line width can represent passenger or cargo volume, while direction can distinguish inbound and outbound movement.
19. Migration Example
Migration flow maps can show sending and receiving regions, but net migration can hide substantial two-way movement. Gross and net maps should not be confused.
20. Trade Example
Trade flows can reveal major bilateral relationships and global hubs. Yet monetary value, physical tonnage and strategic importance are different variables and should not be represented interchangeably.
21. River Example
Hydrological flow maps can follow actual drainage networks and use line width to represent discharge. Here the route is physically constrained and direction is intrinsic.
22. Information Example
Digital communication can be mapped between servers, cities or countries. Geographic distance may matter less than network topology, latency and infrastructure, but the flow still links spatially located nodes.
23. Hostile Test: “That Curved Line Is the Route”
Not necessarily. It may be a desire line showing only origin–destination connection. Always ask whether line geometry represents actual travel path, schematic connection or merely visual separation between overlapping flows.
24. Where Flow-Map Reasoning Breaks
- Line-equals-route: treating schematic connections as actual paths.
- Width blindness: reading line thickness without checking its scaling rule.
- Net-gross confusion: mistaking directional balance for total movement.
- Hairball acceptance: displaying so many flows that structure disappears.
- Threshold invisibility: hiding small flows without stating the cutoff.
- Static-time assumption: combining movements from different periods into one timeless network.
25. Ten Questions for Flow Maps
- What is moving?
- What are the origins?
- What are the destinations?
- Does direction matter?
- What does line width represent?
- Does the line show actual route or connection only?
- Are flows gross or net?
- Has any threshold removed small flows?
- What time period is represented?
- Would aggregation reveal the structure more clearly?
26. Where This Fits
Spatial Interaction owns why flows arise between places. Network Distance owns route-constrained separation. Proportional Symbol Mapping owns magnitude at locations. This article owns the visual encoding of movement between origins and destinations.
The Idea to Keep
A flow map is most useful when it makes the movement between places easier to see than the places themselves.