Scoping our first Data Model (Fast Forward Video)

A summary of the hands-on exercise for data modelers.

  • Example: FDIC Bank Call Report reference data
  • We use five of the FIB Fundamental Supertypes to create a concept map.
  • Transpose the concept map to a Data Model subject area
  • Review the conceptual data model and compare to the ontology graph

The “Fast Forward” version of FIBO data model webinars and presentations is the equivalent to flipping through a deck of slides, the FF has the gist of the underlying education module. The viewer can decide to watch the full video and study the PowerPoint

[advanced_iframe use_shortcode_attributes_only=”true” src=”https://www.youtube.com/embed/pk6Ul0ghnZE” width=”100%” height=”600″ id=”advanced_iframe” ]

You can read the presentation or download the PowerPoint here.

Transcript

Hello and welcome to semantics for data architects “fast-forward”. The Financial Industry Business Data Model, FIB-DM.

In this education module, you would be scoping your first data model from FIB-DM. The fast forward this is an abstract of the class to drive home the key steps in the methodology and to give an overview of what is in the education deck.

About the intended audience and content: So, you are data modeler or data architect. And the education module is a hands-on exercise so you should follow the steps and PowerDesigner or your modeling tool. the example the FDIC Bank Call report and we use fundamental super types to create a concept map, transpose the map to a data model or a diagram, and we compare the model to the ontology graph. So, our use cases is the US Bank Call Report. That’s the report that domestic depository institutions must file with the Federal Deposit Insurance Corporation. And what’s interesting about that report is that the FDIC actually publishes bank filings we can go on the FDIC website and download sample data into MS-Excel.

So, our first step is to draw a simple free from diagram of our sample data. Here in this example we see JP Morgan Chase has a registered address in the United States. They have issued capital. They have a legal entity identifier and they have an FDIC certificate number. Now and then in the the next step we try to replace our circles of sample data with the symbols of the fundamental business concepts.

Yeah and here simply have a list of fundamental concepts for our nine data items and the reasonel for picking them. So, that replaces the circles in our first sample data diagram with a concept map of the identified concepts. And here in this step we still have the sample data in and then the next step is to standardize the relationships, the arrows and here we just draw from the vocabulary of associations and associative entities in FIB-DM. And this is how the final FIB-DM concept map looks like.

We we have a concept map drawn from the  fundamental concepts and we have relationships between the concepts that are from the controlled vocabulary of FIB-DM associations and associative entities. From concept map to creating a data model we follow four simple steps: We populate a diagram with the identified entities. We add the super types of the entities all the way up to the fundamental concept.

With associations and associative entities we can generate an LDM out of our conceptual model subject area. Yeah now the full length video or the or the class does that as a live demo doing the steps in Power Designer. For the fast-forward we just look at screenshots. So we create a new diagram and we pull ins and  entities that made up our our concept map and then we add the super types to the diagram.

So, here as an example we have the Depository Institution. We pull in Financial Institution; we pull in Financial Service Provider; and ultimately we end up at the ultimate supertype our fundamental concepts, a Thing in Role. And we do the same exercise for the Registration Authority, which turns out to be also to be a Service Provider. The stock corporation rolls up to the Autonomous Agent. It’s a Legal Entity. And we have the Country rolling up to Location; and the FDIC registry is an Arrangement. Reference is a catch-all we see the Monetary Mmount rolling up to Reference. The Registered Address is an Address, which in turn is an Index. The Legal Entity Identifier and the FDIC Certificate Number, both are  Organization Identifiers rolling up to Reference. Yeah, and what we should get out of this first step here is our nine base concept from the concept map and hierarchies rolling up to their respective fundamental concepts.

The next step then is to to connect the entities in our diagram and we do that by adding associations and associative entities. And the helping tool is in the FIB-DM spreadsheet, which basically has our fundamental concepts and the participating relationships.

So with the help of that and it was research in the data modeling tool, for instance, we find that the associative entity has Issued Capital connects a Stock Corporation to the Monetary Amount. And the FDIC certificate number is Registered By the Registration Authority and also it is Registered In a Registry. The Registered Address is tied to the Organization via the associative entity has Address. And the Registered Address has a Country.

Yeah, and the identifies Association is a real powerhouse in the FIB-DM model and also in the FIBO ontology. So we have more than  base entities said that participate in identifies associations. So the FDIC Certificate Number identifies the Depository Institution; and the Legal Entity Identifier identifies an Organization and the legal entity. Yeah, and then our final model should look something like this we have our nine base entities and we have seven associations connecting them with each other.

Yeah, and then we’ll do a discussion of the conceptual data model. A question often asked is do I need all these entities? And the simple rule is that in the logical data model after attribution we can just remove entities from the subtype hierarchy that do not have attributes or relationships. And then this is an example how the project logical data model may look like. Here in this case we have only scoped the nine base entities for the LDM.

Denormalizations discussed are to replace associated entities with direct relationships. Well, thanks for watching and as usual you can view the video on YouTube or the FIB-DM website. And there you can also download the PowerPoint of this education module. Thanks again