What is CREST?

CREST is our solution for the rating of consumer loans, mortgages and SME’s. It allows to build predictive models for the estimation of the probability of default (PD) of the counterparties from a set of historical data. CREST is a factory of models rather than a concrete rating model.

CREST allows the storage and management of large amounts of data, provides tools for the analysis of the quality of the data and for the construction of new custom explanatory variables. Through a guided process, users are able to create, manage and validate their rating models based on logit and probit regressions. The selection of variables is eased by several automatic procedures.

Once models are finally built, they can be used in a simple and intuitive way to perform the rating of a set of new counterparties or can be exported to other systems where the rating is to be performed. CREST allows to incorporate qualitative information that affects the final rating and generate executive reports that summarize all the relevant information of the processes.


  • Allows the storage and management of massive amounts of data. The information repositories can be updated with new information, while guaranteeing the replicability of all the calculations.
  • Provides an intuitive and trackable guided proccess, that ranges from data analysis, to explanatory variable definition and model construction.
  • It has multiple automatic tools for assisting the user in the data analysis and the selection and processing of explanatory variables.
  • Incorporates automatic report generation. Scorecards, data reports and model construction summaries can be generated on-the-fly.