I'm using t_positive from Florida_COVID19_Cases (FeatureServer), which is the only documented positive tests column, but as of 2020-04-14 19:00 the total in this column equals 21,629 whereas the dashboard reports 21,628 positive tests, which matches the undocumented t_positive_2 column.

T_positive — Testing: Total positive persons tested for all Florida and non-Florida resident types, including Florida residents tested outside of the state, and those tested at private facilities.

The discrepancy is due to Duval county, and I'm not sure why or what the difference means.

read_csv("https://github.com/gadenbuie/covid19-florida/raw/master/data/covid-19-florida_arcgis_summary.csv") %>% 
    filter(timestamp == "2020-04-15 07:00:37", t_positive != t_positive_2) %>% 
    select(county_1, t_positive, t_positive_2,  c_fl_res, c_not_fl_res, c_fl_res_out)
# # A tibble: 1 x 6
#   county_1 t_positive t_positive_2 c_fl_res c_not_fl_res c_fl_res_out
#   <chr>         <dbl>        <dbl>    <dbl>        <dbl>        <dbl>
# 1 Duval           719          718      704           14            0