Only one more day left of Dashboard Week after this one! It has been a whirlwind of a week, but as I have always been saying in my past blogs, it really has been a blast.


Today’s challenge had us using some NHL data from a website (credit to J Tay for suggesting it to Bethany) to develop a Power BI report to present any stories or insights. Thankfully the data was pretty clean and consistent, so there wasn’t a large amount of cleaning and preparation needed to be done in Alteryx (unlike the last few days!). However, the main challenge with the data was organising it in a structure that makes calculations and analysis easier for Power BI to work with. I built a few macros just to help do these structural changes across all the tables I decided to use. Even though this week is “dashboard week”, I have definitely been able to further develop my Alteryx skills.


Working in Power Query


Despite doing some minor structural changes in Alteryx, I spent most of the time organising the data in Power Query (hey, why not?!). I wanted the challenge to use a tool I am less comfortable in compared to Alteryx. As I was using multiple tables for winners and losers for teams and players, I decided to append the winners and losers tables together, and then pivot the column to a long, narrow structure. The rationale behind this was to use a single measure header to create a specific type of chart (more on that below).


Power BI


For my report, I wanted to focus on uncovering insights as to what make a winning team in the NHL. I wanted to look at game statistics such as goals, shots on target, penalties, total assists, etc. at both a player and team level in order to understand what winning teams actually do compared to losing teams. I wanted to visualise this as a butterfly/mirror/tornado bar chart. Funnily enough though, I have never attempted to build one in Power BI before! However, it turns out all I needed to do was download the Microsoft Tornado custom chart and learn how to set it up. Unfortunately, it seems that this custom chart type is a bit buggy and doesn’t offer the regular features that other charts offer such as filtering and drill throughs. Although this limited the interaction in my chart, I think it was a worthy tradeoff as the tornado chart effectively and efficiently highlights the difference in game statistics between winners and losers.


Final Thoughts


Tune in tomorrow for the final blog of Dashboard Week!



Ben Devries
Author: Ben Devries

Ben graduated with a Bachelor of Music Performance (Honours) from the Sydney Conservatorium of Music in 2023. For the last few years, Ben spent his time working as a professional jazz saxophonist which led him all around the world performing in cities such as London, San Fransisco, and of course, Sydney. But despite his musical background, Ben’s interest in data analytics came from his passion for problem solving and understanding the little details of how and why things work. From there, Ben went on to discover the Data School Down Under, and throughout the interview process became further inspired not only by the logic and flexibility of data, but also the ability for data to provide valuable insights to help solve complex business problems and present meaningful stories. Ben is excited to join Data School Down Under, and hopes to utilise his creativity, improvisational skills, and ability to draw connections upon diverse areas of information learnt as a musician within his new career in data analytics. In his spare time, Ben still enjoys playing his saxophone, as well as downhill longboarding, and spending time with his family.