Kaggle is a platform for data scientists to compete and earn prizes or cash for their insights. Additionally, companies can use Kaggle to have their sets of data interpreted by a vast number of data scientists and they can then select a winning prediction model that is the best one for them. Kaggle offers a win-win scenario for all parties involved. Companies start by establishing a predictive modeling or analytics competition. From there, data miners and statisticians worldwide can compete to produce the very bets models. Kaggle and this crowd-sourcing approach are ideal when it comes to predictive modeling since there are many different ways to come to a valid conclusion. Through Kaggle, companies can review these interpretations and then select the one that is the most accurate, compelling or useful for them. This is much less expensive than hiring on a data scientist full time and could lead to better results since an organization isn't having to rely upon a single person's evaluation or interpretation of the numbers for their future strategies. Also, competitors on Kaggle can team up to offer a more vast knowledge proposition to the companies or they can use Kaggle to interact with one another and get support or "expert" opinions on similar projects. Kaggle is a stat-geek's paradise.Show more screenshots »
Kaggle is a Silicon Valley startup that was founded in April 2010. Its CEO is Anthony Goldbloom who was featured in Forbes Magazine as one of the Top 30 Under 30 executives/ people to watch. Kaggle received a lot of publicity in February 2010 after it received $11.25 million in Series A functions from a venture capital round led by Khosla Ventures and Index Ventures. As of the end of 2011, Kaggle had over 23,00 registered data scientists from around the globe. Organizations such as NASA, Allstate and Wikipedia have employed Kaggle competitions for its predictive modeling needs. Kaggle competitions have resulted in useful analysis for HIV research, traffic forecasting and chess ratings. Kaggle has been featured on BBC, Forbes, The Economist and Financial Times publications.
Kaggle is different from other freelancing kind of sites in that it focuses completely on the science of data- be it predictive modeling, analytics, conclusions, data mining, etc. Those registered to compete on Kaggle typically have an impressive background in this field be it in computer science, statistics, mathematics or other similar studies. Additionally, Kaggle competitors are from all over the world, so there are many unique and interesting models that can be developed thanks to the global insight that Kaggle can offer companies. Kaggle is also different in that it has been quite successful in its short life and has a solid financial backing behind it. Both companies and competitors can assure that this platform is not a fly-by-night and they shouldn't fear investing time and energy into it as they may get a lot out of it.
Kaggle is very well designed and polished. It is easy to start on the site and either register as a company with a competition or task, or to register as a competitor and start submitting reports and conclusions. There are forums available so Kaggle users can ask questions, get more informed and become more acquainted with the Kaggle universe. Kaggle highlights the competitions in a very compelling manner that would incentivize any skilled statistician to pull up their sleeves and get to analyzing the numbers. There are competitions willing to award up to $3 million!
Membership to the Kaggle website is free. It is available to anyone 18 years of age or older. Competitors do not have to pay anything, but competition hosts do need to pay a success fee should their competition result in a successful model.
Kaggle is an awesome place for data scientists to put their skills to the test and possibly make some extra money. Statistics teachers and professors could also use Kaggle as an assignment for their class to engage their students. Nothing motivates more than money! Additionally, companies can save a lot of money and get better results by using a Kaggle competition instead of hiring their own data scientist in-house.