Make CSV processing more robust for USAFacts (or more generally?) #461
Description
USAFacts added a value in the count column which was "1,020" which lead to downstream issues since the column wasn't integers. We could add a preprocessing step (maybe a series of converters for pandas?) that we can apply to clean up these fields. Stripping commas seems to be the main suggestion for now.
Even though this was a USAFacts issue, I think this could make sense to generalize into the utils package.