/Duke-Datathon-2019

🥇1st place out of 57 teams, which included 350 graduate and undergraduate students, in a 11.5-hour dataset challenge sprint. Provided two random forest classification models for Valassis to predict user conversion, depending on the data available to them. AUROC values were 0.64 and 0.86 respectively.

Primary LanguageJupyter Notebook

No issues in this repository yet.