What is Kaggle? For my purpose as a newbie to modeling, it’s a fantastic place to get projects to work on. It’s a platform which hosts data sets for modeling competitions, and it’s really an awesome idea and site to use. When learning a new concept at work, you’ve often got a real problem to work on – the challenge of learning something on your own is that you don’t have this need, and you lack an idea as well as data. This is where Kaggle comes in as an awesome resource – they’ve got data and problems in spades! Plenty for you to learn on.
The really cool thing about Kaggle is how well you’re doing gets scored – you can build your model any way you’d like with the training data set, then score your model via the test data set to see how well you did. You can track your submissions over time to see how your techniques and tweaks resulted in changes (hopefully improvements!) You can also compare your results with friends as well as others across the entire interwebs, as your score is publicly posted in a leaderboard.
This site is going to be virtually all you need for ideas and data to play around with as a newbie. There’s a lot of real deal competitions going on hosted by businesses, as well as learning data sets and projects out there for newbies and experienced practitioners to work on alike. The Kaggle Kernels area is basically a blog area where people share their code, work, and ideas. If you’re stuck improving the accuracy of your model, it is a good place to learn and get ideas from. If you’re looking to play in this area of data, you should without a doubt sign up for a Kaggle account. I plan to use available projects and sets for this site for interesting projects as I work to learn these things myself on this site.