Coding and AI classes in Northampton

The learner downloads the Environment Agency's daily flows for Upton Mill and keeps 81 complete water years, October 1940 to September 2021; a long gap in 2022 is why the run stops there, and four missing days elsewhere are filled by drawing a straight line. The average flow is 1.456 cubic metres per second. The sequent peak algorithm then walks through the days one by one, adding the shortfall whenever the river gives less than the steady amount we want to take, subtracting the surplus when it gives more, and never letting the running total fall below zero. The largest total it ever reaches is the storage a reservoir would need.

The first surprise is the critical period. For a draft of 70 per cent of the average, the worst stretch runs from April 1942 to November 1946, more than four years in which the storage kept falling overall. Asking the program only for the worst single water year gives 22.42 million cubic metres, barely half the true 43.11. Dry years that follow one another are what empty reservoirs, and a year-by-year view cannot see them. At 90 per cent, the critical period runs for more than 23 years.

The second surprise is record length. Using only the last 30 years, a common window for climate averages, the answer at 70 per cent falls to 33.11, because the dry 1940s are left out. Neither number is dishonest, but a design built on the shorter one would have failed in a repeat of the 1940s. The learner tests the function on a tiny invented river where the answer can be worked out on paper, then prints each result with its critical dates. The archive notes treated sewage returns upstream, so the record is not a purely natural river, and our reservoir is imaginary.

Asking how many years sit behind a number is exactly the habit that makes code worth learning for Northampton teenagers in 2026. The longer argument is in why teenagers should still learn to code in 2026.