ID.11. Automatic Profiling in Big Data

Profiling is the activity to detect the significant characteristics associated to a given set of indicators. The objective of this research project is to design a specific system able to extract the data from Big Data ensembles, apply data mining tools to detect the profile of indicators and represent the obtained results graphically for exploitation and dissemination.

Indicators can be characterized from dimensions and other covariates, i.e. the indices of satisfaction, loyalty, etc. may be structured into a hypercube, whose dimensions can be the company, activity branch, country or other geographic unit, date or period of the survey, age, level of education, among others. The objective is the designing of software tools able to deal in a context of Big Data, to automatic exploit to which dimensions and covariates the indicators are related.

To accomplish these tasks we will need to perform a state of the art for automatic characterization of variables, to design the extraction process into an OLAP hypercube of the required information, the specification of the data mining exploitation tools and its implementation and the visual representation of results for its dissemination.

Main Advisor at Universitat Politècnica de Catalunya (UPC)
Co‐advisor at Aalborg Universitet (AAU)