Rejoice! It is football season.
And what better way to win all your friends’ money than by staring at the television and partaking in some good, old-fashion fantasy football! Did you know that with Dundas’ Business Intelligence platform, Dundas BI, you can visualize your fantasy football data on a dashboard (what better way to rub each win in your pals’ faces)? I didn’t at first, but after long hours of research (just kidding), I found a way to make this possible.
This blog will be taking you through a fun tutorial on creating a fantasy football dashboard, which you can then display for the world to see on wall-mounted dashboards, with data from your fantasy league. You’re about to embark on a lesson in unparalleled product flexibility and discover how with Dundas BI, unstructured data is no obstacle.
PS – this isn’t the first time Dundas BI has been used to settle a bet…
This blog will be broken down into a few short steps.
STEP 1 – POWER THE DASHBOARD
First, I need to power the dashboard. And to do so, I need data! Since the 2019-2020 NFL season is relatively young (and I don’t have enough data of my own…yet), to find a suitable dataset, I searched throughout the Internet to find which fantasy app providers make their data accessible (I’m using data from an ESPN fantasy league I found online). I also stumbled upon a few GitHub accounts and forum posts detailing how some Python libraries can be used to pull said data.
This sparked the realization that I wasn’t the only nerd trying to visualize their fantasy league data!
Now that I know I can connect to my data using Python, it’s time to extract my league source data. This forum post provided me details on how to do so. Unfortunately, here’s where the horror began… HOW WAS I SUPPOSED TO WORK WITH THIS? Why couldn’t the data be in a simple Excel file?
My first thoughts were, “Well, back to the drawing board, I guess this can’t be done.” But then I said, “No! I’m up for the challenge!” After a few more hours of online research, and with a little help from the very generous Python language, I discovered a viable solution of how to connect this unstructured data to Dundas BI.
I wrote a simple Python script that would import the JSON data into Dundas BI and give it a more pliable structure. VOILA, I have data – a whole lot of it. Here’s what that script looked like:
STEP 2 – CREATE THE DATA MODEL
Inside Dundas BI’s ETL layer (the Data Cube), I was able to successfully run my script in the Python Data Generator transform, giving me TONS of data to work with. The daunting task of pulling in my data was over. And after adding a few more transforms and reusing my script with different parameters, my data model was complete.
STEP 3 – TOUCHDOWN! CREATE THE FINAL PRODUCT
With the data modeling process finished, my next steps were to 1) put together a few Key Performance Indicators (KPIs) to view on my dashboard to breakdown the results of our 2018-2019 fantasy football season, and 2) make my dashboard as stylish and relevant to its content as possible. After all, what good are analytics if no one’s using them? Here are some great tips on how to design dashboards that’ll have your users cheering for more.
After spilling my creative juices, here’s my final product:
Not too shabby, right? This fun example gives you a taste of how truly flexible and customizable Dundas BI is, and how it can work with any data source or structure. I’m not going to lie: while the end result is ultimately a lesson in product flexibility, this was mostly an opportunity for me to geek out, and I would never pass up that opportunity!
If you liked this blog (and fancy yourself more of a cold-weather sports fan), you’re in for a treat in the not so distant future. On the next episode of Larsen’s blogs, my pièce de résistance, a lesson in 'Forecasting’ and 'AI' in fantasy hockey!
About the Author
Larsen Friis has a passion that lays at the intersection of sports and technology. As a technical resource, he helps build unique solutions for a wide range of clients.