AMS to Ontology: column-mapping cheat sheet
AssociationAI.ai tutorial companion (teaching example, not a case study)

For each column in your AMS export, decide: is it a fact about the member,
or a thing in its own right? Then name the connection in plain words.

COLUMN TYPE          | WHAT IT BECOMES      | EXAMPLE
---------------------+----------------------+-------------------------------
Person attribute     | attribute of Person  | first_name -> "Maria"
                     |                      | email -> maria.santos@example.org
Role attribute       | attribute of Person  | title -> "Safety Manager"
                     |                      | (a title describes the person;
                     |                      |  it never becomes its own node)
Date attribute       | attribute of Person  | member_since -> "2020"
Organization         | entity type          | company -> "Riverbend Freight Co."
Chapter              | entity type          | chapter -> "North Chapter"
Committee            | entity type          | committee -> "Safety Committee"
Certification        | entity type          | certification -> "Certified Logistics Professional"

NAMED RELATIONSHIPS (check the ones that fit your association)
  members work for companies        Person --works for--> Organization
  members belong to chapters        Person --belongs to--> Chapter
  members serve on committees       Person --serves on--> Committee
  members hold certifications       Person --holds--> Certification

THE FIVE QUESTIONS (ask for every column)
  1. What kind of thing does this column name?
  2. Is it a fact about something, or a thing in its own right?
  3. If it is a fact, which thing does it describe?
  4. If it is a thing, what connects it to the member?
  5. What do we call that connection, in words a person would say?

RULES OF THUMB
  - Could the value hold a meeting? Then it is a thing, not a fact.
  - Name every relationship; an unnamed line is a guess, and guesses rot.
  - Leave payment records and note fields out of the export.
  - Confirm every auto-detected guess against the sample values.
  - Regenerate and re-download the schema after any rename.
