Open Conference Systems, STATISTICS AND DATA SCIENCE: NEW CHALLENGES, NEW GENERATIONS

Font Size: 
Interactive machine learning prediction for budget allocation in digital marketing scenarios
Ermelinda Della Valle, Elena Scardovi, Andrea Iacobucci, Edoardo Tignone

Last modified: 2017-05-22

Abstract


Scenario Analysis estimates the relationship between budget allocation and some typical digital analytics metrics quantifying the performances of marketing campaigns. The actual values of the monitored Key Performance Indicators, deriving from different big data sources, are compared to their estimated value after a variation in the investments. Our ensemble approach combines multivariate generalized linear models with machine learning models. R implementation is embedded into an interactive dashboard for providing real time score predictions. The visualization simplifies the business fruition of the analysis and encourages new and deeper experimentations with data.