Now Reading
Are you starting in Power BI? Here’s some best practices!

Are you starting in Power BI? Here’s some best practices!

Power BI, the business intelligence platform that combines an intuitive user experience with advanced analytics, enables organizations to gather data for analysis in seconds and discover detailed insights with built-in AI capabilities.

Are you starting or would you like to start using Power BI?

I leave some tips for those who are starting!

Visuals (Front-End)

  • Tip 1 – Avoid leaving the background blank: The white background does not provide a comfortable view for someone who has prolonged use of the dashboard. I advise you to put a color and adapt it, as much as possible, depending on the company or client.
  • Tip 2 – Create visual logic: Keep dashboards consistent for easier navigation. It maintains this logic and coherence in the titles, filters, tables, chooses to maintain the same organization between pages. When you need to create navigation between pages and/or hide/show visuals, you should use BookMarks.
  • Bonus 1: Less is More!: Avoid saturating the pages with too many visuals/KPI’s, because in addition to being heavy and confusing, the user will not be able to efficiently explore and crunch the information. So I recommend leaving some space between the visuals so that the dashboard can “breathe” and so that the user can analyze the information he wants in isolation.
  • Bonus 2: Storytelling: Every dashboard has a way of being presented and interpreted. The displayed visuals must create a line of reasoning that makes sense to the business. As a rule, this is analyzed from the top left corner to the bottom right corner, making a “Z”, so I recommend that the most important information is arranged on the left, at the top, or at the limit, in the middle.

Data Model (Back-End)

  • Tip 3 – Always use the same date format (especially when it is necessary to merge date fields): This way you avoid data loss and guarantee greater consistency.
    • Tip 4 – Avoid using Many-to-Many relationships: This format becomes highly inefficient as redundant data is very complicated to manage. In the Data Model TAB, if the model has many tables, it helps a lot to create an empty page and drag only the tables you want to analyze.

ADVANCED – When you intend to filter the report by more than one date, you must use a Model of Disconnected Tables (tables without relation).

Nice-To-Have – PowerBI

  • Tip 5 – Have a SQL background: Power BI’s analytical language is DAX and M (DAX being the most important), however most of the time you will be doing metrics in SQL. SQL is very useful for debugging and keeping Power BI efficient, it saves memory and makes Power BI much lighter and faster.

Creating Metrics

See Also
From a Digital to a Cultural Transformation

  • Tip 6 – Use DaxFormatter: DaxFrmatter is a tool created by Power BI professionals that will help you organize the website used to format and organize the DAX code.
  • Tip 7 – Using variables when creating metrics: Mainly when the code is extensive/complex, creating variables is useful to break up the code, making it easier to read and helping to correct errors or bugs. When it is necessary to create a calculated metric/column that uses data from a table that does not have an active connection, the USERELANTIONSHIP function must be used as a CALCULATE parameter. When creating a metric with the SELECTEDVALUE function, the column mentioned in this function does not need to be present in the (visual) matrix to obtain the desired result.

On a last note…

• Tip 8 – Be active in the Power BI Community: The probability that a problem you encounter has already been experienced and solved by other professionals is high, so, finally, my suggestion is that you use Power BI communities interacting with other professionals, clearing doubts, learning from the doubts of others, and even helping others to solve their problems. Attend these communities regularly, even daily or every week and you will definitely be able to help someone and be helped!

Good luck!

What's Your Reaction?
View Comments (0)

Leave a Reply

Your email address will not be published.