We are preparing the Datathon 2020 to take place at 15-17 May this year. This is going to be an somewhat different datathon then the previous ones. It is going to be entirely online. This might make it easier or harder for the participants. Here I will share some of the insights I learned during my participation of the last datathons.
This is subjective to some extent, meaning it might work or not for a different person or a team.
Lets begin with creating a team.
- Create a team with diverse background, meaning if someone does not know the answer another one could.
- At least 1 team member will have to have a larger experience to lead the rest.
- Its a good practice if you already now the members, this will make work easier.
- There might be difficulties if there are people in other time zones, so time management and the division of tasks should be tough out first.
- Not all teams produce results, don’t be discouraged, this is an experience there will be future datathons to attend.
How about choosing the case.
- This depends on your experience and interests.
- Get to know the topic before the datathon, there are tons of examples on the internet.
- Check the sample dataset.
- Research notebooks and papers on the topic.
During the Datathon
- Make a plan.
- Start your article with your main objectives.
- Experiment, this will not hurt. Its better to try then to give up.
- Remember Occam’s razor. The most complex algorithm is not necessary the best.
- Consult with others.
We wish you to have a nice experience. The most important part is not to win but to learn something new and make new acquaintances.