- July 29, 2020
- Posted by: Nancy Wambura
- Category: Power BI
Everyone wants to look good when making a presentation but when your Major was in Finance, your design skills are probably limited. This is where Power BI comes in to not only simplify large sets of data for you but also, help you to effortlessly present data in an aesthetically pleasing way. And just like any other platform, there are some Power BI best user practices that you can employ to make your reports unique and attractive.
Power BI works by using charts and graphs to comprehensively summarize complex data on a dashboard [a graphical user interface that helps in the analysis and visualization of data], and a Power BI report [a multi-perspective view into a dataset, with visuals that represent different findings and insights from that dataset]. Users can create stunning reports using Power BI.
If you’d like to view an example of a Power Bi dashboard, we recently added a COVID-19 Data Update dashboard on the home page of our website. It is an in-house Power BI design that gives you an idea of how your data can look like on a dashboard. It shows COVID-19 updates by country detailing new infections, recoveries and deaths among other beneficial details.
Both new and established users should adhere to the below listed Power BI best user practices when using the data analysis platform not only to better performance, but also, to create stunning reports.
Power BI Best User Practices
1. Stick to the 8-10 visuals rule
Power BI has an array of visuals and when you are a new user, you may assume that you can use any number of visuals when creating a report but that is not the case. Using too many visuals will slow down report performance of Power BI.
It is recommended that users should use a maximum of 8 widget visuals on one report page and a maximum of 10 tiles for dashboards.
2. Avoid Hierarchical Filters
Hierarchy filters are what you use to display multiple category values in a hierarchical tree. Disabling hierarchical filters will increase performance on Power BI. Instead, apply multiple filters for the hierarchy to enjoy better performance.
3. Enable Row-Level Security (RLS)
In our blog post Why You Should Consider Shifting Reporting from Excel to Microsoft Power BI, we covered why row-level security is a crucial feature. We said, “RLS features available for Power BI allow administrators to grant, withdraw or limit access for a team member. RLS has the ability to limit access to particular rows on a report adding an extra layer of security”. This means that Power Bi will display the specific data a user is authorized to view, which improves performance.
However, to get considerable improvements in performance, enable RLS by combining Power BI and backend roles.
4. Manage default visual interactions
All visuals in a specific page fully interact with each other by default. This implies that when a user selects a value in one visual, all the other visuals are filtered or highlighted. This has an impact on the refresh process and depending on the complexity of the calculations on the visuals. It is therefore critical to come up with report designs that minimize the interactions while still achieving the expected outcome for the dashboard user.
5. Import only what you need
When creating a dashboard, data is retrieved from various different data sources then consolidated and key indicators derived in the dashboard. This means that there is a potential to have a lot of detail within the raw data while only high level summaries are finally displayed within the dashboard. It is therefore critical to have the raw data pre-aggregated to some reasonable level using Power Query so as to have only the necessary data imported into Power BI. This not only improves on the Dashboard data refresh process, but also improves on the user refresh time when a user is navigating the dashboard.
6. Keep your Power BI report and data source in the same region
Don’t keep your report and data source in different regions as it will increase network lags. Querying data on Power BI and data transfer will slow down if reports and the data source are in different regions.
Use lighter backgrounds or white ones on reports to ensure you have an easier time when printing.
8. Use template [.PBIT] files
Instead of starting a fresh report especially if you are a beginner, use template [.PBIT] file. Templates can: “1) Be saved with custom colour palettes and themes incorporated in them already, 2) Have corporate branding already applied to pages, 3) Connections to commonly used data sources already in place and 4) Commonly used DAX measures already created” –William Cryger.
9. Use Certified Appsource Visuals
Certified Appsource visuals go through rigorous quality testing by Microsoft. The team at Microsoft ensure that these visuals perform optimally with robust code. Certified visuals are the only ones that can be viewed in ‘Export to PowerPoint’ mode and email subscription.
10. Test custom visuals
For better report performance, test custom visuals. This is because Microsoft does not emphasize on testing custom visuals that are not certified. In cases where you are working with large data sets, some of the untested visuals may work poorly. When you discover this, change the custom visual to one that performs better.
11. Work with filters more than slicers
While filters and slicers work almost the same way, they have some differences. Slicers works by generating two queries. One query will get the data and the other will fetch selection details. This means that when you create too many slicers especially when working with large data sets, performance will slow down.
Filters, which show specific data to users, are better to use in this case but have a major weakness in that when working with large data sets it has slow loading time. However, it’s certainly better to user filters over slicers.
The above are some of the best practices you should observe when using Power BI. Always remember to consider how to communicate to your audience. Make it easy for the people who will read the report or those you are presenting to, to understand the data.