Abstract:

Big Data often has demographic and behavioral data, but little to no attitudinal information. This presentation describes the many challenges in using survey data to add attitudinal information to Big Data. We also describe our solution to align demographic and behavioural data stored in large client databases with attitudinal data collected in a survey that was used for developing consumer segments with different needs. We will show that higher predictive validity for survey data does not always mean higher predictive validity for Big Data, and that data fusion and augmentation can be a helpful solution.