Input: a line seed=S resamples=B, then a CSV district,pm25 with two districts.
Let a and b be the readings of the alphabetically first and second district. With rng = np.random.default_rng(S), repeat B times: resample a and b with replacement (rng.choice(x, x.size), a first then b) and record median(a*) − median(b*). The 95 % CI is the 2.5th and 97.5th percentiles of these differences (np.percentile). Print Lucheng median 46, Ouhai median 38, difference 8, bootstrap 95% CI of the difference (2000 resamples, seed 42): [3, 13.5] (values in :g format) and the CI excludes 0 or the CI includes 0: the difference may be due to chance.