Five Key Steps to Quality Data Modelling

To most business people data modelling is a black art that is practiced by the company IT department. It brings no tangible business benefits and is solely designed to make mere mortals feel confused and inferior.  Sadly, many IT departments actually foster this view and, for those that do, the data modelling that they carry out fails to deliver any real benefits, in spite of the mumbo jumbo they chant about it.
It does not have to be this way. Data modelling, when done properly, can deliver enormous business benefits to any enterprise, that include:

• Higher quality information for all business activities
• Easier access to that information
• Robust information systems
• Better identification of products, profit and cost centres.
• Elimination of redundant and unneeded information
• Reduced costs and increased revenues

But how do you do data analysis and modelling “properly”?  Where do you start?
Here are two simple rules that will guide you on your way:

Rule 1: Extract all information for data modelling from exactly the same sources from which you extracted your Business Functions.

Rule 2: Only model data that is needed to directly support the Business Functions of the enterprise.

Starting with Rule 1 will make sure that you conform to Rule 2.

The Integrated Modelling Method provides a foolproof technique for extracting candidate entities, attributes and associations from the sources from which business functions were extracted.  This technique can be used by novice and experienced analysts alike.

These sources include:

• Transcripts of taped analysis interviews with senior business managers.
• Typed up notes of supplementary information from these interviews.
• Function titles and descriptions developed during function modelling.
• Information flow diagrams produced in analysis workshops.

Technique

The basic technique:

Example

The first steps in the technique are best demonstrated by way of an example.

The following is a transcript of an interview with a business manager:

“We receive orders for our products from our customers the day before they require delivery. We check the quantity of the raw materials required to bake the products and, if necessary, we order more from our suppliers. We bake our products fresh each morning. We make deliveries to our customers several times each day.  At the end of each week we invoice each customer for the deliveries made to them during the week.  We accept payment or remittance from customers by cash and cheque only”.

All of the noun structures have been underlined.

The first sentence is:

“We receive orders for our products from our customers the day* before they require delivery.

Working through the sentence one underlined noun at a time we get the following list of candidate entities and associations:

order <means of ordering> product
product <ordered by means of> order
order <received from> customer
customer <the source of> order
product <delivered by means of> delivery
delivery <means of delivering> product
customer <recipient of> delivery
delivery <made to> customer

Note: day* is an attribute of order

Because every association is two-way, when we document one association we immediately create its reverse.

Working through the whole of the transcript gives us the flowing complete list (sorted alphabetically):

This short extract has given us eleven unique candidate entities and thirty (15 x 2) candidate associations.

Rationalising Entities

The list of candidate entities now needs a little more work to remove false of spurious entities.  A typical example of a candidate item that is not a proper entity is “invoice”.  An invoice is probably the most common business item to be incorrectly modelled as an entity.  The invoice itself is a piece of paper that represents a business entity or a collection of entities such as a sale (of one or more products) or a billing (for one or more sales).  These are the actual data entities that should be modelled - not the pieces of paper representing them.

Converting Associations to Relationships

The associations that we identified must now also be rationalised and converted to “Relationships”.  Associations simply tell us that two entities are associated and give us a suggested name for that association.  A relationship tells us all of the essential information that we need to know about the association.  This includes

• the precise name of the relationship
• whether it is mandatory or optional
• its “degree”, that is, is the relationship is one-to-one, one-to-many or many-to-many

The Entity Relationship Diagram

All of the preceding information is essential to know but is almost impossible to visualise and of limited use without constructing an Entity Relationship Diagram (ERD). This is the single most powerful model for use in the understanding of the data structure of the enterprise and an essential element in the design of quality databases.

The many ends of a relationship on an ERD in the Integrated Modelling Method are indicated by the following symbol , which, for obvious reasons, is called a “crows foot”. If we turn this symbol upside down we get a “dead crow” .  This gives rise to one of the most powerful, yet simple, rules for achieving a really effective layout for any ERD, which is “Dead Crows Fly East“. This means that all “many” ends of relationships in ERDs should appear either like this or like this on the diagram.

The net result of this is that all high volume and volatile entities will appear to the top and to the left of the ERD and all low volume and more constant entities will appear to the bottom and to the right, as in the following simple ERD.

All of the rules which were briefly described in this post are expanded on in full detail, together with many more, in the Integrated Modelling Method eBook: Data Structure Modelling, available at our Online Store.