Getting ready for your first datathon
Some thoughts to ease your journey
A datathon is a time-bound competition to create data-driven solutions to address a business case with data provided and leveraging your creativity and analytics skillset.
Why participate in a datathon?
Based on my experience, a datathon is an effective way to polish and build up your analytics skills and contribute to something meaningful with the added benefit of working with data from a different industry than your everyday work. I recommend participating in a datathon even if you don’t have extensive data analysis skills because the journey helps you figure out how to approach a business problem and being able to use data to create solutions to address the problem. You might not be successful the first time around, but sticking around and learning from others participating is also a great opportunity. Also, it is quite mind-blowing to see at the end how different everyone’s solutions are even though everyone is trying to address the same broader problem.
How do I feel ready?
First — Grab your team. Most datathons typically allow 4–6 members on each team. When you are looking for team members, make sure you get individuals that have different experiences and skillsets than yours. A mix of data analysts, data scientists, user experience designers, front-end developers, back-end developers, project managers, business analysts, etc. usually makes up a good team. If you don’t have a team, don’t worry since most datathons allow you to pair up with other individuals who are also participating without a team.
Second — Clearly define the problem statement. Typically datathons release the theme ahead of the competition but the theme usually isn’t fleshed out enough to start coming up with solutions. Exploring the theme and the dataset(s) provided can help you refine the theme to begin defining the specific problem statement you will be addressing. Clearly defining the problem statement will help you and your team to stay on track and finish with a solution.
Third — Understand the submission requirements and judging criteria. Some datathons require a working solution at the end while some don’t. Taking a minute to identify what’s needed and how your solution will be judged will help you identify the focus area while you build out the solution. Typically solutions are pitched via a 5-minute presentation using no more than 5–7 slides and are judged based on scalability, creativity, impact, scope, and relevance.
Fourth — Identify additional open datasets to help build your solution. Datathons typically come with at least a dataset that you need to leverage to build out your solution. Bringing in additional open datasets can help your solution come across as more creative and will enable you to harness more insights from the data. And more relevant data is usually never a bad thing.
Fifth — Know when to stop and focus on the presentation. Datathon solutions are typically presented to judges and audiences after 36–48 hours. Your solution is judged based on how you present it and make it relevant to the theme and to the judges so taking the time to craft your message and your presentation is as important as creating the solution. The presentation has to be a good mix of showcasing the technical capabilities of the solution as well as advertising a simple solution that a layperson can understand. Balance is the key!
Sixth — Take advantage of the SMEs available to you. Subject Matter Experts are made available to you to ask questions and get clarifications. You have the data experts as well as the people who will be impacted by the solution available to ask questions and understand the problem at hand better based on their experiences. Take advantage of the opportunity and don’t develop your solution in a silo without engaging with the end-user.
Seventh — Don’t compromise on data ethics. Maintain the integrity of the data and don’t try to fit the data to your solution. If data tells you a different story than what was envisioned, update the messaging to reflect what the data is telling you. Never compromise on data ethics to make your solution seem better.
Eighth — Don’t forget to network, learn, and have fun. Connect with your team and other teams. Be open to learning from others. Most importantly, don’t forget to enjoy it! It’s stressful but it’s a great experience to learn new things and challenge yourself.
You got me excited! How can I participate in one?
Women in Data Science is hosting a virtual datathon on Kaggle that runs from January thru February and you can sign up here.
Other ones I recommend are — NASA Space Apps, The Opportunity Project (this one is typically months long), and Virginia Datathon.
