The shape of ten thousand runs

Every run you record is, at heart, a list of points — latitude, longitude, a timestamp. On its own it's a thin thread: out the door, around the park, home again. Beautiful to you, illegible to anyone else.

But threads pile up. Lay enough of them over the same streets and something happens that no single run can show. The thread becomes fabric. The fabric has a shape. And the shape tells you things about a place that no map ever could.

What you're looking at

The field above isn't a photograph or a drawn map. It's rasterised — every point is plotted with a little additive glow, so wherever runs overlap, the density builds up and brightens. Quiet lanes stay dim; the popular loop along the river burns bright. It's the same idea behind tools like Datashader: don't draw the data, draw how much data is there.

Scroll back up and watch the threshold — points don't all arrive at once. They rain in, so the shape assembles the way it does in real life: a few early risers, then the commuters, then the whole crowd.

Why this matters for ontrack

The point of ontrack is to turn your runs into numbers that actually help. Most of that is personal — your bests, your training load, your trend. But the data also has a collective shape, and that shape is genuinely useful:

• Where it's safe and social to run at 6am in a city you've just moved to.
• Which segments are contested enough to be worth racing.
• How a parkrun course is really run, lap by lap, by thousands of feet.

This is a field note — a sketch, not a product. But it's the direction the lab is pointed: take the data we're already collecting, and find the pictures hiding inside it.

A run is a sentence. Enough of them, and you can read the city.