The geekiest geeks solve the trickiest challenges from across the world.
Imagine trying to work out how to disaggregate the energy consumption of each domestic appliance in an apartment block. Such a challenge would be beyond the skills of the average real estate developer, appliance company or apartment owner. Yet the solution could help all three make better decisions about how to moderate energy usage. Such a problem might become solvable if it could be put to a community of the world’s leading data scientists, so they could compete, for fun, prestige and prize money, to find a solution. Kaggle is bringing together that global community: some of the geekiest geeks ever.
Kaggle’s community of more than 110,000 data scientists is made up of academics, consultants and employees working in their spare time. They come from the fields of econometrics, predictive modelling, artificial intelligence and statistics. Teams compete to solve tricky problems posed by clients. Thirty-six teams, for example, competed to win $250,000 by coming up with an algorithm that General Electric wanted, to help airline pilots avoid turbulent weather.
The 110 competitions run on the platform in its first three years attracted hundreds of teams from all over the world. The most popular, a contest to predict who would have survived the sinking of the Titanic, attracted more than 6,000 entries.
Scientists who do well earn points that boost their ranking in the community. In August 2013, the leader was Leustagos, a computer scientist based in the Brazilian city of Belo Horizonte. The top 30 contributors come from, among other places, the US, Singapore, Hong Kong, the Russian Federation, Slovakia, India and Spain. Companies who want access to this elite global talent pool can do so through Kaggle Connect, which matches clients and challenges to consultants.
Kaggle is a small but potent example of how global communities of talent can be brought together to address challenges, which might otherwise prove elusive.
Image courtesy of infocux Technologies
Last updated: 09th of May, 2014