On Time, Everytime

March 6, 2019 3:27 PM

"When a train starts running from one station to the next station, conceptually, these two stations will temporarily be closer to each other. And that is exactly what this visualization shows: whenever a train moves to the next station — and only for as long as a train is moving — the origin station moves towards the destination station. The faster the train, the closer it moves to its destination. " - Jan Willem creates a shifting, flowing, almost-alive rail map of The Netherlands.

via FlowingData

posted by the man of twists and turns (25 comments total)

The animations are beautiful and mesmerizing. But the explanation feels like it’s missing something. Surely a station doesn’t move. Stations don’t physically get closer to each other, right? I’ve never seen a train station physically “get closer” to any other point in space as a dynamic effect.

So...is he referring to a relativistic effect of space-time distortion? Or is he referring to a center-of-mass for a station’s trains?

I’m puzzled at what he means by “the stations get closer”.
posted by darkstar at 3:41 PM on March 6, 2019 [2 favorites]

I hate math problems like this.
posted by Fizz at 3:51 PM on March 6, 2019 [1 favorite]

On re-reading, I note he uses the word “conceptually”, but that still doesn’t quite clear up the concept he is visualizing here...
posted by darkstar at 3:57 PM on March 6, 2019

The stations aren't moving iRL, they're just moving in this visualization.

Watching them move is sort of telling you something about how active each station is; the major ones are whipping around like crazy, while the ones out on the edges just kind of twitch one way or the other every now and then.
posted by egypturnash at 3:58 PM on March 6, 2019

This is really cool and reminds me of a short story hook I've had forever, but never know(n) what to do with. Inspired to finally write it out and save it away in my folder of bad ideas.
posted by I'm always feeling, Blue at 3:59 PM on March 6, 2019

The only way I made sense of it is that the stations get closer to each other from the perspective of the train... but then, the station you've just left is moving away from you at exactly the same rate as the station you're moving towards is moving towards you. So it doesn't actually make any sense. Plus stations have to go multiverse if trains are approaching from two different directions.

I wonder if it's related to a time travel game I play with myself on the London Underground. Say I'm waiting at station A for a train that's going north. That next train is approaching station C, two stations to the south, and is ten minutes away from me.

The travel time between A and B, the intermediate station, is four minutes, between B and C is five minutes. So the train I want is a minute away from C.

A southbound train on the same line arrives at my station A. It's going to B. If I catch it, then it takes me four minutes to get to B, by which time the train I actually want is two minutes away from B, approaching from the south. I get off the southbound train at B, look at the arrivals board and see that the train I want is now just those two minutes away.

Four minutes ago, it was ten minutes away.

It feels as if I've jumped four minutes into the future, and I've had a free train ride into the bargain.

Aye, we make our own fun in London.
posted by Devonian at 4:10 PM on March 6, 2019 [6 favorites]

It might be easier to figure this out if we knew exactly what open datasets were being used, rather than just "open datasets from NDOV Loket", because there are like a thousand of them.

I guess if there is like a "baseline efficiency of travel" between say Eindhoven and Tilburg of whatever, 30 minutes, but three trains run ten minutes late, it distorts the baseline which impacts the conceptual distance.
posted by turbid dahlia at 4:21 PM on March 6, 2019 [1 favorite]

Oh for Pete’s sake...I was completely misinterpreting what he was saying in the explanation.

My original interpretation: so, he’s saying that, in some way, the actual stations are getting closer in real life. This must be metaphorical, or figurative, or something. So, like, the center of mass of all of a station’s trains shifts in a particular direction, maybe? Then his animation visualizes that movement. But what in actuality is moving, that is then being visualized here?

My new interpretation: no, you dummy. He’s just saying that he made an animation that shows stations artificially moving around, based entirely on train movement.

Adoy.
posted by darkstar at 4:37 PM on March 6, 2019

From a passenger's point of view, for me it makes sense conceptually if two stations get further away when a train leaves from one to the other. A station is "closer" to my current location if there is a train leaving for it in the near future. You could visualize this as the stations creeping closer and closer as the departure time nears (starting from a baseline of say an hour beforehand) and then suddenly snapping apart when the train leaves the station, as a destination effectively gets further away the moment you miss a train.
posted by L.P. Hatecraft at 4:39 PM on March 6, 2019 [12 favorites]

It's cool looking but I also don't get it.
posted by runcibleshaw at 4:50 PM on March 6, 2019 [1 favorite]

If a train leaves one station and heads to another, he artificially moves the first station’s position so that it is closer to the second station.

Thus, the animation shows train movements, visualized artificially as station movements.
posted by darkstar at 4:55 PM on March 6, 2019

I think the intuition here is that two stations are "closer" if there are fast, frequent trains between them. Like Amsterdam and Haarlem are "close" because there's a fast train between them every few minutes. Some random rural station is much further from Amsterdam because there's only one slow train an hour. The part that confuses this a bit is the dynamic visualization; I'd understand a static map based on this train frequency metric better.

It's conceptually similar to an isochrone map, only the metric is not just "how long does it take to get to point X" but more of a squishy "how convenient is it to hop a train to point X".
posted by Nelson at 5:09 PM on March 6, 2019 [3 favorites]

I feel as L.P. Hatecraft does that I am close to the next town when I *can* get on a train, not when I've just missed one. Though I am not sure that switching the algorithm would change the simulation much, it might just change the clock labels.
posted by clew at 5:29 PM on March 6, 2019 [1 favorite]

If stations Y and Z are an equal distance from station X, then if a train goes to station Y every five minutes and station Z every hour, then station Y is temporally "closer" to station X... on average, you'll get there more quickly than station Z when you randomly show up at station X. You might get lucky on a single trip to station Z if you time it right and show up right before the train there departs but 4 minutes before the trip to station Y departs, but over hundreds of trips, the difference in temporal distance would become apparent.

(I have no idea if that is what the creator means, they in fact say "temporarily" and not "temporally")
posted by lefty lucky cat at 5:51 PM on March 6, 2019 [1 favorite]

It helps if you turn off the "distortion" switch.
posted by jeremias at 6:01 PM on March 6, 2019

When the train is in motion, the stations are neither dead nor alive, they become Schrödinger's stations
posted by greenhornet at 6:44 PM on March 6, 2019 [2 favorites]

I'm from MetaFilter and I could overthink a plate of beans a train map.
posted by ricochet biscuit at 6:46 PM on March 6, 2019 [2 favorites]

It helps if you turn off the "distortion" switch

just don't forget to turn it on again
posted by flabdablet at 7:30 PM on March 6, 2019 [1 favorite]

Cool. It shows you how many trains are moving in a certain direction at any given time, by moving the station into that direction. So, for example, it becomes really obvious that the overall motion in the morning is towards the West (de Randstad) where a lot of jobs are, and in the evenings it's the other way around.
posted by monospace at 10:31 PM on March 6, 2019 [1 favorite]

Think of the line between two stations as the time between them.

When the train is moving between them, the time is short. When there is no train moving, it's going to take you longer to get there, so the line is longer.
posted by rokusan at 10:49 PM on March 6, 2019

Its quickens, in the old sense of the word, the representation of the network. Very nice.
posted by stonepharisee at 11:46 PM on March 6, 2019 [1 favorite]

Let's just be honest. This data visualization does not aid understanding of anything. It's just shiny.
posted by dsword at 11:46 PM on March 6, 2019 [5 favorites]

The nation as amoeba...
posted by Devonian at 3:44 AM on March 7, 2019 [1 favorite]

But, if I’m walking through all of the carriages of the train on my way to the station it’s definitely closer at some stage and then the same distance at the end (unless I walk back to the end, in which case it’s further).
posted by unliteral at 3:48 AM on March 7, 2019

It takes a little bit of work to understand the point of this, but after staring for a while I got how the distortion communicates the extent and direction of connectedness. But I had to stare at Rotterdam for a while, toggle distortion, and realize where it actually is - and then realize that the "pulled in a direction" is essentially communicating the extent of connectedness with other areas. It's a neat demo but there are probably better ways to calculate and communicate this.
posted by entropone at 6:24 AM on March 7, 2019