What’s the implication of assigning “Identifier” vs. “Organization” as the semantic type of a data element? Will the data element be used differently or interpreted differently?
Hello Christine,
First of all Welcome to the QuickSight community !
Directly from the best practices -
When you create a topic using automatic data prep, Q will automatically select relevant semantic types that it can detect. Q uses semantic types to understand which specific fields to use to answer vague question like who, where, when, and how many
For example, in the student enrollment statistics example, Q already set Home of Origin as Location so if someone asks “where,” Q knows to use this field (Figure 6). Another example is adding Person for the Student Name and Professor fields so Q knows what fields to use when your business users ask for “who.”
Another important semantic type is the Identifier . This tells Q what to count when your business users ask questions for example “How many were enrolled in biology in the year ?
Here is a list of semantic types that map to implicit question phrases:
Location: Where?PersonorOrganization: Who?- If there are no person or organization fields, then Q will use the identifier
Identifier: How many? What is the number of?Duration: How long?Date Part: When?Age: How old?Distance: How far?
Note : Semantic types also help the model in several other ways, including mapping terms like “most expensive” or “cheapest” to Currency . There is not always a relevant semantic type, so it’s okay to leave those empty
Fore more info see - Best practices for enabling business users to answer questions about data using natural language in Amazon QuickSight
Hope this helps.
Cheers,
Deep
2 Likes