Building your portfolio with Kaggle

Kaggle’s claim to be the “home of data science” has to be taken into perspective. As we have discussed at length, Kaggle is open to everyone willing to compete to figure out the best models in predictive tasks according to a given evaluation metric.

There are no restrictions based on where you are in the world, your education, or your proficiency in predictive modeling. Sometimes there are also competitions that are not predictive in nature, for instance, reinforcement learning competitions, algorithmic challenges, and analytical contests that accommodate a larger audience than just data scientists. However, making the best predictions from data according to a metric is the core purpose of Kaggle competitions.

Real-world data science, instead, has many facets. First, your priority is to solve problems, and the metric for scoring your model is simply a more or less exact measurement of how well it solves the problem...