Use correct Pandas Datatypes for columns

[olejandro]

Currently rows containing na values are not dropped before generating output (i.e. csv files). However, some of the outputs (the ones not filtered by attribute) are based on larger tables (e.g FI_Comm) from which all rows and only some columns are used for a particular csv. This results in a large number of additional rows generated by the tool. Could we add a switch that would control whether the rows containing na values are dropped? At this stage, it is probably best to drop the rows by default.

[olejandro]

@samwebster @siddharth-krishna

[siddharth-krishna]

#63 was a temporary hack to filter out rows with None string values. Let's keep this issue open to track progress on using the correct dtypes for columns so that we can just use pd.dropna. This will also help us avoid having to explicitly convert the Years column to ints, etc.

[olejandro]

#66 applies the hack to all the columns.