Coding and AI classes in Batley, West Yorkshire

The learner centres a square 4 km on each side on the point that OS Open Names, served through postcodes.io, gives for Batley, then marks a grid of 41 by 41 points 100 m apart: 1,681 points in all. An open elevation model returns the height of the ground at each one. The lowest reading is 46.7 m and the highest 157.8 m; the mean is 105.2 m and the median 107.7 m. Between neighbouring points the ground rises or falls by 4.8% on average, and the steepest single step on the grid is 24.7%.

Sort the heights and the curve draws itself: for each level, the fraction of points at or above it. To squeeze the whole curve into one number, measure where the mean sits inside the range, (mean minus lowest) divided by (highest minus lowest). That ratio is the hypsometric integral, and for the Batley square it comes to 0.527. The ground is spread fairly evenly through its range, leaning slightly high: 56.9% of the points lie in the upper half of the range, yet only 2.0% reach its top tenth, so the high ground is broad shoulders rather than sharp peaks.

Now the trap. The integral depends on the single highest and single lowest points, and a small sample rarely contains either. We drew 10 of the 1,681 points at random, 2,000 times over. The highest height seen averaged 138.9 m instead of 157.8 m, and the integral wandered between 0.407 and 0.661 across the middle 90% of draws. With 100 points it stayed between 0.487 and 0.592, and with 400 points between 0.507 and 0.553. A regular grid 1 km apart, 25 points again, gave 0.543 and a highest point of 155.2 m. More data narrows the answer, but the extremes are the last thing a sample finds.

A Batley teenager who has watched a summary swing like that will question the next confident figure an AI hands over. That instinct comes from writing and rerunning the code, and it is a sound reason to keep learning to code in 2026. The longer argument is in why teenagers should still learn to code in 2026.