GOOAAALLL: Hacking our way to predicting World Cup qualifiers
Which teams will qualify for the 2022 FIFA World Cup?
In late September 2021, students and professional data scientists came together for a datathon (or data hackathon) to solve this critical, worldwide question. ⚽️ 📈
The 5-day datathon was jointly organised between Atlassian Data Teams and the University of New South Wales (UNSW) Data Science Society (DataSoc). It attracted 100+ students and Atlassian data professionals, who teamed-up and got “stuck in” learning the game of “soccer.”
The students competed to create the best predictive model of qualifying teams, or the best infographic that would inspire sponsorship.
The results were fantastic, and of course we’ll share the winning team’s predictions…but first let’s dive into some post-game analysis and highlights from the field.
Hackathons are not just for coders
Atlassian has been organising hackathons for a looong time. We run company-wide 24-hour hackathons, called ShipIt, and they are a big (and fun) part of our culture. Sometimes they even result in new products and innovations, like our Jira Service Management product.
We’ve long known that the hackathon projects that make the biggest impact come from teams with diverse talents — design along with data, project management along with coding — and this proved true again for this datathon.
Fan Jiang, an Atlassian Data Scientist mentoring the winning team, observed the following:
What really stood out to me was some of the groups organised themselves to have very different skillsets and strengths so they complimented each other. The winning team organised themselves to have a solid programmer, a strong modeller, strong visualisation person, and someone who was good at presentation/thinking about the business problem.
Unfortunately though, the common perception amongst students is you need to be a good programmer to join a hackathon. A lack of confidence in one’s coding ability is a key reason behind low hackathon participation, explained Ivan Hou and Vincent Chen, vice presidents of the UNSW DataSoc at the time of the hackathon. And many data science students don’t get exposed to coding until late into their college education. They’ll just sit out hackathons early in their degree and miss out on valuable experiences.
We wanted to encourage anyone to participate. So our datathon incorporated a beginner-friendly infographic challenge that rewarded diverse skills and out of the box thinking. This dramatically increased the participation and submission rate of beginner teams.
Organisations can further break down this “coders-only” perception by creating more specialised events for non-coder skill-sets. For instance, some hackathons cater specifically to data, design and product management professionals. Another great idea from the UNSW DataSoc is to sponsor new-comer friendly “Data Starter-thons.”
Remember: hackathons are not just for coders.
Both staff and students get a lot out of hackathons.
As to why some students attend hackathons, coding skills or no coding skills, Josh Amoils, one of the winning students offered the following thought:
A datathon is the most raw reflection of a data scientist’s day-to-day that a student can experience at university. You get thrown in a pressure cooker and very quickly figure out if you hate it or love it!
We found students were new to many of the tasks involved in practicing data science in the real world. They needed lots of help understanding the business problem, and hands-on guidance in cleaning raw data sets. The hackathon gave them an opportunity to get this experience and be better prepared for their first job.
And even though volunteering at a hackathon can be hard work, staff at Atlassian love getting involved. Nina Kumagai, Data Scientist at Atlassian, had the following to share:
I would recommend every analyst at Atlassian to try and take part in this event as a volunteer next year. There’s always a lot to gain and learn from these students even as an established data scientist/analyst.
So… which teams will make the World Cup?
The winning team came up with the following prediction:
Will this play out? As of publication (Mar 2022), with 15 of the 32 slots decided, this prediction has gotten it right 11 out of 15 times. We won’t know the final performance until June 2022. Our fingers are still crossed for the Socceroos. 🇦🇺 🤞
One thing is for certain — both students and Atlassians have a deeper appreciation for the beautiful game.
Thanks to the following people for their help with the article:
Josh Amoils, Vincent Chen, and Ivan Hou, from UNSW;
Shahin Elliin and Sarah Eade from Atlassian.
Where can you learn about the next hackathon?
- If you’re a student, your uni or college may be organising one — e.g. the UNSW Data Science Society, for students of UNSW
- Hackathons Australia (AU) and Major League Hacking (US) also list up-coming events
Finally, if you’re a student interested in joining our team, we have internships and graduate positions frequently posted in our careers page.
.
