Dissecting a Datathon: It's About Community and Mindset

In looking back at Alberta’s largest datathon event to date, not only are we compelled by the ideas and solutions that were explored that weekend – but also the human factor. 

It was an impressive goal - making life better for people - but the audacious objective served as a rallying cry to bring Alberta’s best minds together to collaborate. 

Over two days, 250 data scientists and AI/ML developers huddled into 45 teams to crack open data sets, identify issues worthy of tackling, and test solutions. Some participants were from as far away as Russia and South America. They spoke multiple languages, came from varying backgrounds and expertise, and brought differing opinions and perspectives … all with the intent of solving real, human problems for Alberta and the world. 

If we look at the data of datathons, the metrics we’d typically dive into include the number of teams, the amount of data, the number of failed models vs. successful ones and, of course, the validity of the solutions proposed. And we’ll do that (stay tuned for interviews with winning teams), but in addition to the six projects that resulted from the event, one of the metrics that proved most important was a far less tangible one: mindset.

Wherein developing advanced tech solutions usually means working in secret and holding on to IP with white knuckles, the ATB datathon was a means of supporting open innovation and creating community connections across students, researchers, and developers from both private and public industry. 

The datathon forced teams needed to think about more than data. Criteria looked at viability and market fit. Would the proposed solutions in fact be feasible? Would the market adopt them? Would they have a real impact? 

And a time crunch helped. The limited time to create data models and test projects meant all teams - no matter if they were made up from perfect strangers or long-time colleagues - needed to be creative, adaptive, and cut failures loose ... fast.  

Much like the work done in ATB’s Leap Innovation Lab (where we imagine, experiment, build and launch new products/solutions that rethink finance from the ground up)  teams competing in the datathon faced a number of constraints that determined the success of the solution as they began the development process. 

Developing new solutions to challenging, complex and ambiguous problems is never easy. In fact, a few teams dropped out of the competition before the end of day two because their models or ideas simply didn’t pan out. This is something that we are all too aware of in the technology space - sometimes solutions just don’t work, but failure is part of the process. Even some of the most talented technologists can face roadblocks. Sometimes a model seems like it could work in theory, but fails in integration. 

Though failing to advance can feel like a blow to a team during the event, it’s not. We’re all the better for it. The benefit of a datathon event model goes beyond the success of a single idea to focus on community, connections, and shared learnings. Failures become valuable lessons for everyone involved – as participants learn from one another and tap into experiences from industry partners (in this case, IBM, SAS, GDG Cloud Calgary, Axxiome and AMII) who lent their tools and subject matter experts to teams (in advance of the event and during).

Connecting humans is one thing. 
Cracking open and providing access to data is another.  

Datathons like this show how critical open data is to building meaningful solutions. While most of the data sets provided at the datathon were public, we also created synthetic datasets – computer generated data that simulates authentic data but protects the actual data it was based upon – so that the teams could develop real solutions for real challenges. This allowed us to continue protecting the privacy of ATB customers, while still working on innovation that could benefit real people. It’s vital for organizations to contribute public and/or synthetic data to open datasets so that there is a greater datapool to create solutions for urgent issues.

Looking Forward...
Ingraining agile development methods across all of us who aim to solve problems (no matter how big or small) is key to collective success. ATB was proud to have played a role alongside the community and our partners to bring so many people together to prove out how rich a resource data is; and how technology can be a means of making us better humans. 

We plan to host and co-host more datathons and community events to continue to strengthen Alberta’s innovation mindset and continue the momentum that’s building world-class AI/ML capacity here in our own backyard.

ATB is working on digital transformation for a new future. Want to know more about what our Leap Innovation Lab is building? Check us out today.

We are ATB transformation - innovating at the forefront
of robotics, AI, blockchain and the future.