In Figure 4 , the effect of age by hour is more pronounced. Older riders outnumber younger riders for the bulk of the daytime, but that edge decreases dramatically by 8 P.M. From 9 P.M. to 4 A.M., younger riders slightly outnumber older riders.
Method
The test was performed using the chi-squared test of proportions, which suits the nature of the two proportions being compared in the null hypothesis and also aligns with the methodology introduced in the HW4 notebook reproducing the results from the recidivism study \cite{repository}. It is useful in cases of testing with a binary categorical variable (in this case, day vs. night) and two population groups (young vs. old) to compare the observed proportions to the expected proportions.