Add ability to set dtypes or converters

[aiguofer]

Currently, when we pull data with df_to_sheet, every column gets set with dtype 'object'. It'd be nice to have a way to provide column dtypes or converters for each column similar to pd.read_csv.

See: https://stackoverflow.com/questions/47507426/pandas-dataframe-filter-by-dateform-that-doesnt-have-dtype