A trajectory drawn the usual way
One person carried a GPS logger for 24 hours. This is how the data is usually shown: the fixes joined into a line on a blank map.
The data
Columns: time, x_km, y_km. The line uses only the last two.
Where did the day start, which way did the person go, how many places did they visit and for how long, was any of it fast? None of that can be read from the line. A line has no direction, no order and no duration. The small scribbles hint at stops, but only because GPS jitters while the person stands still.
Add time: order, direction, stops
Same fixes, coloured by time of day. Press play to replay the day. The clock shows when.
What time gives us
- Direction and order. Home in the north-west first, then south into downtown, east to the mall, then a wide loop around the east and south side before ending in the south-west.
- Stops. A stop is where the clock keeps running but the position does not. A simple rule finds them: all fixes within m for at least min.
Detected stops
We now know when and for how long. We still cannot say what the person was doing at each stop.
The space-time path
Time on the vertical axis. This is Hägerstrand's space-time path: the whole day in one picture, with no animation needed.
Reading the path
- A vertical segment is a stop: position fixed, time passing. The two long ones are home (before 10:00) and the south-west corner (after 12:06).
- A sloped segment is movement. Slope is speed: steep means slow, flat means fast.
- Look at the segment just before 11:50. It covers the widest loop on the map in almost no height. Something happened there.
The 2D line in Step 1 is this path seen from directly above. Every vertical segment collapses to a dot and every slope is lost.
Speed: derived, not measured
Nobody recorded speed. It is distance divided by time between fixes, smoothed over a s window so GPS jitter does not look like walking.
Mode of travel from speed
What speed reveals
- Every leg is by car at a steady 35 to 40 km/h: ordinary city driving.
- At 11:38 the person leaves the mall and heads south along the eastern arterial at the same 40 km/h. Nothing unusual.
- At 11:45 the speed jumps to km/h and stays above 100 for a 10 km loop around the south of the city. Same car, same roads, a completely different driver: this is someone being chased.
- The speed drops to zero on a residential street in the south-west and stays there 10 minutes. Then a short, steady 40 km/h leg, and nothing for the rest of the day.
- Speed alone splits one drive into two episodes, a normal drive and a chase. On the map they are one line.
Spatial context: what is at each stop?
Add the land-use map and the points of interest. The stops are the same; now they have names.
Stops with context
A 17-minute stop at a café is ordinary. Twenty-six minutes in a jewellery store and 43 in an electronics store, followed by a drive that turns into a 150 km/h chase, is a different story. The chase ends with a 10-minute stop on a residential street with no shop, no café, nothing there to explain it. Then twelve hours at the police station.
Context explains five of the six stops. The sixth is explained by what comes before it and after it: a chase, then the police station. The pawn shop is on the map, but the track never reaches it. Where the driver was heading is a guess, not an observation.
From a line to a story
Stops and moves, in order, with a place for every stop and a mode for every move. This is a semantic trajectory.
The day, reconstructed from the fixes
What each layer added
| Layer | What it told us |
|---|---|
| Geometry only | Shape and extent. Nothing else. |
| + time | Start, end, direction, order, and where the person stopped and for how long. |
| + speed (derived) | Mode of travel, and one leg that does not fit: a normal drive that becomes a 150 km/h chase at 11:45. |
| + spatial context | What five of the six stops were. The sixth is explained by the sequence around it. The stops became activities and the day became a story. |