This repo contains the winning solution to the Cajamar PythonHack 2016 Predictive Modelling challenge, part of the activities of the PyConES 2016 conference held in Almería, Spain, on 7 to 9 October, 2016.
The actual submitted solution (described in Spanish) is hidden in the repo's history. The current version has been translated into English, includes some cosmetic changes and fixed typos but it is functionally identical to the submitted solution.
The interested reader can find additional details (in Spanish) about the PythonHack 2016 Datathon in the official site.
The overall contest consisted of two categories: Credit Card Analytics and Predictive Modelling. The source code in this repo is only related to the latter. A detailed description of this challenge (again in Spanish) can be found here.
According to the terms and conditions of the contest, we are not allowed to publicly disclose any dataset. Sorry for any inconvenience.