My first Hackathon 🏆
Analysis of a dataset containing broadcast data provided by R&S. Predict future behavior of the data in order to set up a plan to optimize both user experience and revenue simultaneously. An ensemble learner highly dependent on tree based models was used, followed by a custom greedy optimization. The Exploratory Data Analysis {tidyverse}, the modeling {caret} and the optimization was implemented in .
During the finals in Munich with the best teams of the online competition worldwide the solution was presented in front of and evaluated by an expert jury. Finally we achieved one of the top ten final scores.
Note: After now having participated in a couple of hackathons I have to give big Kudos to everyone that was involved in the organization of this amazing international celebration of engineering! It really was perfectly thought through and executed from the main problem to be solved to all the subtle details throughout the finals!