Trends in Power BI – my observation after two rounds of the dashboard duel.
In the rapidly evolving world of data analysis and visualization, Power BI has become an essential tool for businesses. Last year, the Dutch Power BI Dashboard challenge “Dashboard Duel” was launched. Participants can utilize a dataset provided by a charity, develop an impressive dashboard, and the charity actually adopts the winning dashboard. The challenge contributes to the adoption of Power BI, even among organizations reliant on donations and thus having a limited budget. Additionally, it serves as a valuable learning experience for developers eager to practice and experiment with real-life scenarios and data. A win-win situation!
As a member of the jury evaluating the dashboards, I have the pleasure of closely observing how participants utilize Power BI to create good, usable dashboards. My assessment focuses on the “back end,” namely the data model, consisting of (Power) Queries, Power Query setup, DAX, relationships, data types, application of best practices, and so forth. It’s a fulfilling role to fulfill!
After two rounds of judging, I noticed several trends in Power BI, which I’m delighted to share in this blog.
The flexibility of Power BI: even a flat file can work
The datasets for the duel are often quite limited in nature, typically consisting of just one table. What strikes me with every dashboard I review is how many possibilities there are with such a simple, flat table. While a “star schema” is ultimately the best variant of a data model, I found it inspiring to see the beautiful things that can be built on a single table without much additional data modeling. And for me, that’s also proof of the flexibility of Power BI. You can develop true enterprise data models with it, but the tool is also perfectly suited for building more basic dashboards based on such flat files.
Field parameters are becoming more and more commonplace
It was an exploration that took me deep into the core of the company. I learned about their financial processes, their customers, their services, and much more. Each new insightful discovery brought me closer to creating valuable dashboards that would truly enhance the client’s business operations.
Attention for the datamodel
I always evaluate the top 10 submissions, and it always strikes me that the data models of those dashboards are quite tidy. Measures are used, neatly organized in folders, the DAX is usually formatted properly, and even VARs are utilized. All of these are positives that I didn’t necessarily expect, so it’s great to see. And it’s also super important because a well-organized data model works better and is easier to maintain.
Documentation remains a concern.
Was it all positive then? No, it wasn’t. It struck me that careful documentation could still be significantly improved. Documenting the model, the functioning of the dashboard, the measures, and everything that happens in Power Query; it’s important but happens too infrequently.
And perhaps that’s somewhat understandable. In Power BI, documentation is often still primarily a manual action, which makes it labor- and time-intensive. There are various external tools available that offer more insight and thereby facilitate the process, but there’s still a lot of manual work involved. And these external tools are not native Power BI features, are also not officially supported by Microsoft, so you’re reliant on permission and sometimes even assistance from your IT department to install and use these tools. That certainly doesn’t help.
Conclusion
The dashboards I’ve had the privilege to evaluate provided me with a fascinating insight into the setup and the level of others. And I’ve seen some impressive developments! The Achilles’ heel (documentation) seems to be a challenge for others as well, which is “fortunately” reassuring. But I have confidence that Microsoft, along with us as the Power BI community, will ultimately find a good solution.
I’ll also be judging in the next rounds of the Dashboard Duel, and I’ll certainly share the interesting findings once again!
Steven Annegarn
King Expert Officer @ Data Kingdom
