The antibody test results for this data set don't seem representative of the population at large. Reason being, if you look at the antibody positivity rate, it consistently drops as time goes on (whereas everything I've read suggests it should increase with time as more ppl have been infected).
My guess is that at the beginning, hospitals were only testing COVID-sick patients, and as time went on testing became more widespread.
Or the initial antibodies drop in people that have had COVID over time (which they probably do). The antibody tests test for the kind of antibodies that are present immediately- over time a previously infected person will have less of those antibodies and more of the back up Tcell kind. You can read a review on this here: https://www.nature.com/articles/d41586-020-02400-7.