This is a competition for Data Science Retreat program 2016 based on the following Kaggle Challenge:
Predict Grant Applications
This task requires participants to predict the outcome of grant applications for the University of Melbourne.
https://www.kaggle.com/c/unimelb
We were divided into two teams, one with the students on the Data Science track and another with the students on the Data Engineer track. We had only 3 days for accomplish this challenge.
Our solution consisted on creating a hierarchical model that combined features based on data of this 3 big groups found in the dataset
- People model
- Team model
- Grant model
As we are a multidisciplinar team, with different backgrounds (engineering, economics and finance), our expertise in programming language varies, thus, in order to maximize our throughput, each one used the language in which he felt more confortable (Python and R).
This is the result we got in just 3 days:
If you want to learn how we achieved it, click here