How to define the adequate cash prize sizing for hosting a Kaggle or similar compeition?

0

If you take Kaggle as a well known example of data science competition size, how do you know what is an adequate budget for the cash prize size? At least to determine an order of magnitude, given I am not able to study all previous competitions, cluster them by this factor and assess driving factors? irony - could be a competition itself?

I've also looked up the Q&A page at Kaggle and have found more or less same question but without answer yet:

We would like to understand:

  • What is the minimum prize pool that starts capturing your attention?

..

  • Open-ended question - could you give us idea on what are the most important things you take into account before deciding to participate are not?

Open Food Broker

Posted 2019-12-12T09:25:07.303

Reputation: 361

Answers

1

This is a generic question that may need more than just cash prize.

1) Organizational Reputation 2) Dataset Size 3) Defining Problem statements with Domain knowledge 4) Time period 5) Computation required for the contest. 6) How curated/clean is the dataset and much more.

If the problem statement is huge and the reward is in single digits, its not worth anyones time and at the same time, if the problem statement is mediocre and huge money is rewarded, it does not serve the purpose of the Kaggle competition at all.

even though Kaggle provides GPU and 16GB ram with CPU's, sometime the Kernel crashes and there is no work around unless you spend some money on AWS/Azure instances.

I hope this gives few pointers

Syenix

Posted 2019-12-12T09:25:07.303

Reputation: 339

thank you, @Syenix! do you have to provide a data set or can it be a part of competition, or can be a compeition limited to delivering a best data set sized by requirements? in terms, they say that big part of data science is collecting and cleansing data. I see the point with the hardware factor where for bigger problems you take investments risk. – Open Food Broker – 2019-12-12T10:26:56.867

1You can use the datasets as long as it complies with terms and conditions. you will get the T&C' when you signup for the competition. Most Datasets are proprietory to the owner. Its always best to provide your own dataset for the competition. And again when you upload to Kaggle, Kaggle has the rights on the datasets which you should read and also get it reivewed with your legal team if its related to actual customers. – Syenix – 2019-12-12T10:31:06.957