Online coding and Python classes in Kingswood

The learner requests heights from the OpenTopoData service for a 48 by 40 grid, about 6.2 km by 4.4 km, around Kingswood: 1,920 cells ranging from 7.8 m to 114.4 m, with a median of 53.3 m. A linear stretch maps the lowest cell to black and the highest to white. But the land is not spread evenly: split the height range into ten equal slices and the middle two hold 391 and 353 cells, while the top slice holds only 90. Equalisation replaces each shade with its position in the cumulative count, using NumPy, so a shade that sits above half the picture becomes a middle grey.

With a linear stretch nearly half the picture, 46.5%, crowds into one quarter of the grey scale, covering heights from 34.4 m to 61.1 m. After equalisation that band of greys covers only 39.8 m to 53.1 m, so small differences in the busy middle ground become visible, and the average grey difference between neighbouring cells rises from 9.52 to 12.2. Nothing comes free, though. The lightest quarter now has to cover everything from 69.9 m to 114.4 m, so detail on the highest ground is squashed, and the picture uses 187 grey levels instead of 251 because some rare shades merge. Equalisation redistributes contrast; it does not create information.

A Kingswood teenager who has written equalisation by hand will ask what any AI filter threw away, and Python is where that question starts. The longer argument is in why teenagers should still learn to code in 2026.