I basically agree with JoelBrown, but let me elaborate a little.
In the IT world, as JoelBrown says, a "data type" is a technical term referring to a specific format of data, like "32-bit integer" or "date stored as the number of days since Jan 1, 1970". Or arguably in a slightly more general sense to refer to a category of data that can be used in specified ways in a given programming language or other software product, like "string" or "boolean". (The difference I am getting at here is that we may not care exactly how, say, a boolean is stored, whether it is a single bit, an integer where zero means false and non-zero means true, the character T or F, etc. We just care how it works in the program.)
"Type of data" would be read by most IT people as a general phrase whose exact meaning must be determined by context, rather than as a technical term with a well-established meaning.
Thus, if you wanted to say that your programming language supports both 32 and 64 bit integers, you should refer to these as "data types". But if you want to distinguish how your software product uses data imported from other systems versus how it uses data entered directly by the user, you should call these "two types of data".
(It suddenly occurs to me that your question may have nothing to do with IT. If you are talking about different kinds of data in a statistical analysis with no reference to whether that data was collected and processed on a computer or manually, then nothing I've said applies. :-)