Introduction
Last week in our Power BI class we focused on one of the most important concepts in data analysis: data modeling. Data Modeling is the process of analyzing and defining different data type that is collected or produced. I learnt that before creating dashboards or visualizations, it’s critical to properly structure and connect your data. Our teacher illustrated the importance of using power query to remove duplicates, handle blanks , filter and sort data and merge and reshape datasets.
Joins vs Relationships
Joins are used in tools like SQL or Power Query to physically combine tables into one. In Power BI, joins usually happen in Power Query before the data is loaded into the model.
Relationships in Power BI link tables without merging them, allowing for more flexible and efficient analysis.
we were taught that Power BI prefers relationships over joins because they help maintain clean, scalable data models.
SQL join types
An inner join returns only the rows that have matching values in two tables.
A left join keeps all the rows from the left table and only brings in matching rows from the right table.
A right join does the opposite—it keeps all rows from the right table and matches what it can from the left.
A full outer join returns all rows from two tables, whether there’s a match or not.
A left anti join returns rows from the left table that don’t have a match in the right.
Similarly, a right anti join shows rows from the right table that don’t have a match in the left.
Types of Relationships
We explored different types of relationships:
One-to-Many is the most common (e.g., one product category to many products)
Many-to-One is the reverse of one-to-many
Many-to-Many is more complex and is used when both tables contain duplicates
I also learned about cross-filter direction, which determines how data flows between tables.
Data Schemas
Another important concept was data schemas, especially:
Star Schema
One central fact table
Connected to multiple dimension tables
Best practice for Power BI models
Snowflake Schema
More normalized (dimension tables split further)
Can be more complex but sometimes necessary
The star schema stood out as the most efficient and recommended approach.
Practical Application
During the session, I practiced:
Creating relationships between tables in Power BI
Defining primary and foreign keys
Avoiding common issues like circular relationships
Structuring datasets into a star schema format
Personal Reflection
Before this class, i knew power BI as a tool for only presenting data. Now i can use it for cleaning data and creating visuals.I understand how to connect data and that how data is connected is just as important as the data itself.
Top comments (0)