Title: Extracting insights from large and complex datasets: Examples of dimensionality reduction by applying economic theory Authors:  Tsvetomira Tsenova - Experian Bulgaria (Bulgaria) [presenting]
Abstract: Currently, an increased number of datasets containing individual microdata from surveys and regulatory reports of banks become publicly available, which increases the information universe for academics, policymakers and the general public. However, insights generation is insufficient due to the dataset's volume, complex dimensionality and changing structure over time. The lack of an adequately long and consistent time series structure hinders purely empirical research explorations. The purpose is to provide several examples of how economic theory could be used to enrich micro-data sets with additional statistical data and focus on certain dimensions for answering specific policy and general public questions. The examples include the survey of professional forecasters for the Euro Area and the United States, as well as EU regulatory bank balance sheet data. The insights relate to monitoring the state of inflation expectations, economic growth prospects, structural uncertainty, lending decisions, credit risk and financial stability.