By Precision Sourcing
As you may have noticed Precision Sourcing hosted 50 data analysts at Facebooks offices last Friday with the goal of delivering innovative insights from a rich set of NRL data. Now at this stage, I must apologise as we rushed out a newsletter last Thursday and grammatically it was a big fail, but this is what happens when you get super excited about an epic event such as the NRL DataJam.
This week Precision Sourcing was lucky enough to catch up with the winning team captain Danny Peppou, who presented magnificently against some very, very tough competition (at this stage I must mention Quantium who came a very close second).
Below is what Danny had to say about the Precision Sourcing NRL DataJam.
1) What made you enter the 2017 Precision Sourcing NRL DataJam?
I have a passion for the intersection between analytics and business application and saw the NRL Datajam has the perfect blend between the two. As an entrant to the 2016 Datajam competition, I knew what I was in for – large data set, many avenues of exploration, high-quality competition, scrutiny from my peers and a great day of fun, food and footy!
2) How did you approach such a rich set of data, what was your strategy?
Given that we had limited time to work on the business problem we needed to isolate a problem that we knew could be explored in such a short period. We went straight to the game transactional data, as it had many similarities to many other customer analytical datasets (think banking data, retail data, etc.have) we had worked with before that made it easy to see what kind of models we could build. We then brainstormed what kind of things we could accurately predict given such a rich data set and arrived at our player benchmarking business problem.
3) What were the key insights you feel impressed and won it for you and the team?
After we built our model we scored all of the hookers and halfbacks. From this we showed a graph ranking the players with their benchmark score. Our NRL Analyst (Leigh from the Rabbitohs, who was our NRL SME for the day) knew we were on the money when he looked at the top ranked players – the ranking and the score between players was very accurate to his perception of the game. At that point we knew we had to tell this same story to the judging panel.
4) How was it presenting and being questioned by such a knowledgeable sporting panel?
I was very nervous being judged by 3 NRL greats, Brett, Greg and Des, who combined had both the on field experience and the game analytical experience. They definitely brought a high level of credibility to the competition. I knew that any ideas we presented would need to stand-up to scrutiny. Although their questions were tough, it gave me confidence in our idea and how we had communicated it – they understood our concept and were probing deeper into it. I really enjoyed that part of the experience.
5) From the other presentations what stood out for you as “damn it I wish we had done that” moments?
We definitely missed an opportunity to visually bring our idea to life with more visuals. We could have come up with a nice visual dashboard showing the players and their ratings, which their performance across matches. This would have shown the depth to our idea. However I think focussing on the quality of our analysis meant that we ended up with something very valid/reliable – even if we didn’t showcase it as well as we could have.
6) Any insights from the other teams you feel were particularly interesting?
I really liked the player engagement model KPMG presented which showed where NRL players were coming from, from all the lower leagues. I think that the feeder / pathways programs within footy are crucial for its long term success, so understanding where the leagues strengths and weaknesses are is crucial. This is something I would be keen to see developed further to aid in the recruitment of younger players and fans.
I was also interested in how a couple of the teams focused their analysis on the probability of combined players winning. The teams that went down this avenue showed it yielded some very valid results. I think if we could combine our individual player benchmarking analysis, and roll this up to how different players played together on a team and versus certain competition (at a player / team level), we would really be on to something even bigger and better.
7) How did you celebrate the win?
I called Dario (my other teammate on the ground with me), who was about to board the plane back to Canberra. He didn’t believe me when I told him we had won – it took a photo of the prizes to convince him that we had actually won. We were both in awe of the depth and polish of the other presentations. My inbox exploded on Monday morning from all of the congratulations from my colleagues and other team members. We still havn’t fully celebrated our win – I think we are saving it up for State of Origin Game 2!
To get involved in the next Precision Sourcing DataJam contact Joel Stein.