When importing data from a csv file the number of rows checked to determine the data type is limited to 100 rows. There are cases where the first few hundred rows are set to null before a result can be seen. In such cases even if the result is of type Integer Spotfire sets the data type to String. When there are 1000+ column in a data table it becomes tedious to change the data type manually. Please submit a ticket to TIBCO requesting to add an option to scan whole column when determining the data type.
Below is an example of how this is implemented in another statistical software (JMP).