Visual Predictive Analysis for Business Data

Traditionally, company decisions have been made based on a few indicator variables, business intuition or pastexperience in the field. However, in our current digital era, there is more and more access to data coming frommultiple sources that companies cannot ignore at the time of making decisions. Yet many companies lack thenecessary analytical expertise to take advantage of this data abundance. Some of the reasons behind thisdifficulty includes not only the size of the data, but also being able to integrate data from different sources,different levels of trustworthiness, and coming at a high speed. All these factors contribute to make dataunderstanding and modeling a complex process. Many companies from the public and private sector are facingthese types of problems when they need to take decisions based on data. Management consultancy firms facethe additional challenge of working with many different companies with related but not identical problems.One of the main objectives of this research grant is develop predictive analytic methods that allow stakeholdersto understand the data and be able to adjust the model to different scenarios based on their data knowledge. Inthis type of analytical approach the transparency or intuitiveness of the model is important, so that the analystcan understand what the path from data to results is. In this context, interactive visualizations are used as themeans between the computational model and decision makers.