9 - Conceptual Database Modelling and ERDs

Table of Contents

Reading Materials

Welcome to Week 8

This week we dive into the conceptual components of designing a database using the data modelling process and creating a UML Crow's Foot Entity Relationship Diagram (ERD). By following a detailed series of modelling steps, the student will be able to design a basic database eliminating many to many relationships and establishing relationships and their types.

Conceptual Data Modelling

Data modelling is the process of following a series of pre-defined steps to conceptually design a database schema and then developing a graphical representation of a proposed database design using entity relationship diagrams. This design will be based on the discovery of required entities, business rules, and database design requirements.

8 Steps of Data Modelling

If searching on the Internet, you will find many versions of the data modelling process steps. The following version is a simple and effective methodology that follows the principles of this course and will server you well in designing database schemas.

Step 1 - Identify Entities

Identify the major objects of which we need to store. These will likely be people, places, events about which the client needs to store data. It is easy to lose focus on this first step and start thinking attributes (or characteristics) are entities. You must be clear in determining what is an entity vs an attribute.

For Example: For a used car lot, some entities might be: Vehicles, Staff, Locations, Services offered. Be careful not to chose something like type of car as that is an attribute and not necessarily an entity. It is almost always more beneficial to start with less entities than more.

Step 2 - Identify Attributes

Identify the characteristics, or properties, of the entities that need to be stored.

Examples: Vehicle Colour, Staff First Name, Staff Last Name, Vehicle Make, Vehicle Model, etc

In addition to the attribute names, determine if the attribute is mandatory or optional data. Mandatory data are included in bold format.

Step 3 - Specify Unique Identifiers (PKs)

Define attributes within each entity that could be used as a unique identifier for individual instances of the entity. This will ultimately become the primary key, and therefore must meet the criteria for a primary key. If a suitable one could not be identified, then a new surrogate key can be added to auto-generate a key field. Primary Keys are indicated in the diagram by being underlined. They will always be bold as well as the unique identifiers will always be required.

Example: Social Insurance Number (SIN), Vehicle Identification Number (VIN), productSerialNumber, etc.

Step 4 - Define Attributes Which Form Relationships

Identify entities that have relationships with each other. At this point it does not have to be tied to a specific attribute, although it ultimately will be.


Step 5 - Define Relationship Cardinality / Optionality

Determine the number of occurrences of one entity for a single occurrence of the related entity. This typically will only have three options - "none", "one" or "more than one". More than one is known as "many". Each relationship cardinality is defined by a range.
i.e. none or one, one or more, none or many, one and only one.
These are the possible results and are determined in BOTH directions of the relationship. (i.e. # entity 1 for each entity 2 and # entity 2 for each entity 1)
Be careful not to write these based on the existing data, but based on data that may occur in the future as well.
Lastly in this step, the Key attributes are identified as Foreign Keys that relate to their related entities Primary Key.


Step 6 - Eliminate Many to Many Relationships

If the cardinalities for any one relationship contains a many in both directions, then it must be eliminated. This is because it is physically impossible to create a many-to-many relationship in a standard relational database.

To eliminate the many-to-many relationship, erase the relationship completely and create a new entity between the two previous entities (called a junction or bridge entity). The recreate 2 new relationships between the new entity and one each to the two original entities. In most, if not all, cases, these new relationships will be opposite 1-to-many relationships.


In each example above, then the new entity is created, the designer must go all the way back to step 1 from the new entity and follow the steps forward to ensure complete design and consistency.

Step 7 - Normalize the Database

The normalization process is the largest step, by far, and involves determining dependencies between attributes, and their unique identifiers. The next 3 weeks of this course covers this large topic and then you wil understand that the data modelling process via ERDs overlaps with normalization here and you need to know both in order to perfect your database designs.

The main goals for normalization is to:

Step 8 - Define Attribute Types and Sizes

This final step is simply to define the attribute data types and sizes. These values are not required for any of the previous steps and therefore it is recommended to do this last to avoid having to redo them several times. Some things to note:

Entity Relationship Diagrams

ERDs are a graphical representation of the proposed database showing the entities, attributes or characteristics of the entities, and the relationships between the entities (people, places, concepts, and events).
There are 4 Goals for the ERDs:

Symbology and Formatting

In this course, we will focus on the CROW's Foot style of UML ERD. The standards for a UML Crow's Foot ERD are:

The cardinalities of the diagram are displayed using only 3 symbols.

EVERY relationship line will have TWO symbols on each side representing the range of values.
There are only 4 possible combinations that are used.

In the above example:

a couple more examples to learn from:

Using Diagrams.net (Draw.io) to Create a Crow's Foot ERD

The link to the referred to website is https://app.diagrams.net/

There are other online tools or you can use MicroSoft Visio to complete the ERDs. This site will demo draw.io as it is fairly simple, free and produced clean ERDs.

Creating a New ERD

When you first visit the site, you are prompted to either create a new diagram or open a new one. Your ERDs can be saved to your local harddrive and reopened later for further editing.

Click on Create New Diagram and then choose a crow's foot diagram. This is the last option under the software category. Name your diagram and click Create.

You will now have a new template diagram in front of you that you can break apart and uses the pieces to create your ERD. Make sure you expand the "Entity Relationship" object window (as shown below) to have access to additional objects you may need.

You will also want to open the format panel to edit the fonts, sizes, relationship symbols etc. Open the panel by clicking the middle icon in the top right corner. This panel is context sensitive, meaning it will changed depending on what object is currently selected. To edit cardinality symbols, make sure you have a relationship selected.

Note the difference between the entities provided in the sample diagram. There are samples of both standard single field PK entities and bridge entities with a composite key identifier.

Saving Your Work

Once you are ready to save your work, click the file menu (the one inside the browser window, not the browser file menu) and chose Save. Indicate the location you want the file saved, make sure it has a good name and click save.

The next time you need to work on the diagram, you can load the website homepage and choose open and existing diagram and navigate to the file you just saved.

IMPORTANT: The save file from this website or other websites will NOT be appropriate for submitting your work. You should first ensure your name, the date and your id is in a box on the diagram and then take a PNG screen shot or save as a PDF document for submission.