Take a normalized data model of the project data and convert it to a dimensional data model. develop a dimensional model diagram of your facts and dimensions. Remember, some data is heterogeneous, necessitating that it is moved into its own fact table (see target information below).
Download the normalized model of the source system.
Download the dimensional model template.
Open the normalized model in draw.io.
Open the dimensional model template—save it as ProjectDimensionalModel.drawio.png.
Using the [login to view URL] file, create your dimensional model—export it as a .png file when you are done.
The data model should include:
Designed as a Star or Snowflake schema.
Names for all the tables, specifically identifying which are dimensions and which are facts.
Names and data types for all the columns in your tables (make sure the data types are visible and not hidden).
Use of Surrogate Keys.
Primary Key identification.
Foreign Key identification.
Inclusion of Natural (source system) keys.
Relationships lines to logically connect the keys between tables.
Commentary on each table on what it is for, what the "grain" (one row in the table) is, and why you modeled it how you did.
Acceptable ways to do this are a Word document, or in plain text in [login to view URL], but NOT putting each comment into the properties of each table in [login to view URL]; we have trouble viewing those often.**
Targets (CSV files)
Incorporate the targets by designing any necessary table(s) to hold them (heterogeneous). Notice they are annual targets, which is common in the real world, but not very useful for any analysis that is not for the entire year, so you will store daily targets (instead of annual) in your model. Ensure you consider this when you design your table(s).
Target Data - Channel Reseller - Store Download Target Data - Channel Reseller - Store
TargetData - Product Download TargetData - Product