Update data_us to fix MS data import issue

[tgehrs]

Mississippi has a similar issue to Wyoming where much of their data is excluded from your model because the positive is often equal to the total. It looks like once per week they do a mass update of their total number of tests.

Currently only days are shown in your chart/model when a mass update in total test count occurs.

By adding MS to the zero_filter it should show the accurate data for Mississippi for all days and therefore reflect in the model/charts.