Tips for Data Science Competitions

Interested in trying out a data science competition? Learn why you should give it a go and get some tips for your next challenge.

Why should someone participate in a data science competition?

Data science competitions provide the opportunity to learn new skills and techniques, as well as improve existing ones. They keep geoscientists sharp and allow them to win prizes and build up a portfolio of work. Competitions are fun, challenging, and the perfect way to hone one’s skills or get involved in the industry.

Xeek is Studio X’s platform for data science challenges. On Xeek, participants join teams to solve challenges, bring to life diverse ideas, and compete for a chance to win rewarding cash prizes. Xeek challenges unite the data and geoscience communities around the shared goal of crowdsourcing innovative solutions that solve the biggest challenges in exploration.

What are some tips for data science competitions?

Get Started Early 

Keep an eye out for challenge announcements and dig in early. For some competitions, the number of daily prediction submissions is limited. So, it benefits you to get started early on to give yourself all the time you need. 

Collaborate

For Xeek challenges, you can participate in teams. Working in a team can help you come up with new creative ideas, get over roadblocks, and it's often more fun! Check the Challenge Terms for the challenge you’re considering to see more information on how to work on a team.

Attend Office Hours

The Xeek team will usually host office hours for challenges. This is your opportunity to learn about the challenge from the creators and ask any questions you may have. Take advantage of this time — it may give you the spark you need.

Just Do It!

There’s no harm in downloading the data and giving it a go. Just jump in and see where things take you. You’ve got nothing to lose, and so much more to gain.

How are Xeek challenges assessed? What sets winners apart?

Success in a Xeek challenge looks different for each one. For some challenges, winners are assessed on a purely quantitative level — the best performing algorithm when tested against a private data set wins. For other challenges, winners are selected based on more qualitative measurements. For example, in the “Keep Geoscience Weird”challenge, solutions were assessed based on the reproducibility of their code, the quality of documentation, the complexity of data visualized, and the quality of the visualization to select our winners. 

Xeek challenge winners come from all over the world. They have a wide range of disciplines and experiences — there have been winners in their early 20s and winners in their 80s; those with PhDs in geoscience and those with no geoscience experience. Anyone can be successful on Xeek!

Does Xeek have any data science-focused challenges coming up?

The Indirect from the Source Challenge is a good place to get started. There is no prize for this challenge, but it helps you get started working on Xeek. The starter notebook walks you through how to download data, clean data, build a machine learning model, and submit machine learning predictions to Xeek. Xeek launches challenges periodically throughout the year, so keep an eye out for data science challenges in the future.