This project develops and evaluates predictive models to identify museum card holders likely to churn. It proposes an optimized retention strategy combining phone calls and emails, prioritizing high-value customers and maximizing profit within a €5,000 budget constraint.
View Report »
The following work is the result of a university project in customer analytics, during our MSc in Data Analytics for Business & Society at Ca' Foscari University of Venice. The authors are Gabriele Bidoia, Giada Pezzo, Beatrice Fabris, Giacomo Sarrocco.
Since it involved the use of sensitive customer and business data, we are unfortunately unable to publish the entire dataset and several parts of the report had to be removed. In particular:
- The original data files are not present in the github repository
- The references to specific museums or locations are deleted
- Information about customers, such as age, geographical distribution and so on are deleted
- We are still working on releasing
data.R, where we cleaned and preprocessed the data
However, the core of the project, consisting of the chapter on churn prediction and retention analysis, is published in its entirety.