Implementing Nowcasting Techniques for STS publications

ongoing

The aim of the action is to develop and implement a modelling process to provide earlier and reliable estimates of the named STS publications, that can be disseminated as experimental statistics to be timelier in communicating important changes about the construction industry to the public and policymakers. The modelling process will focus on applying Machine Learning Techniques for estimating two STS indicators: the Construction Prudcers Price Index and the Volume Index of Construction Production. There are several although not significant risks in the project that are related to: 1. the available data sources, 2. the efficiency of the ML methods. Within the scope of the project there will be a thorough examination of available data sources outside and inside the HCSO. Especially we plan to use administrative, public and private sources to ensure the quality of input data for the ML techniques as it is crucial to have the appropriate data and not only the state-of-the-art approach to process it, and for processing we are building on our previous project that created the base for the nowcasting procedure. We plan to further develop and update the nowcasting toolkit.