Acquisition modeling in the nonprofit sector based on Facebook data

This research investigates whether Facebook data can be used to predict donation behavior. On the basis of individual specific Facebook data (i.e. Facebook groups, pages, and page categories), we build an acquisition model that predicts an individual’s likelihood to become a donor. Three different classification algorithms are being compared, namely Bagging, Random Forest and Adaboost. Furthermore, by constructing different models, we assess the value of different Facebook features as predictors of donation behavior.

Publication year:2019