SAS for Finance

Data-gathering exercise

Since the team wanted to leverage internal data, a data quality exercise needed to be conducted. Multiple variables were probably going to be needed to predict CPI. These variables were in turn classified into various categories. The categories were based on areas where the modelers thought the consumers were primarily spending money. The hypothesis was that, by using these categories, they would be able to identify significant categories that predict CPI. The lower the number of categories that would come out as predictors, the less the team would have to invest in ensuring that the data quality remains consistent with implementing the model. The initial list of categories of interest that the team had are as follows: ...