In this repository, you can find the files, notebooks, and the related dataset of the lessons on Data-driven building behaviour prediction and simulation which have been offered in the context of "Energy and Environmental Technologies for Building System" course at Politecnico di Milano (Energy Eng. Program)
Regarding the first case, the scripts are a simplified version of a part of implementations that my M.Sc student Manoj Manivannan had developed in his thesis project. The dataset which is employed in this project is the AC consumption data of a residential Building in Austin Texas. The complete results of this study can be found in this paper.
The dsetata corresponding to only one building is provided in this repository (considering the Pecan St. Inc' license agreement); in order to have access to the device consumption of several buildings provided by Pecan Street, you should create an account(free for Academic use) on DataPort.
The second case (classification of commercial buildings -- in progress) is a simplified version of the notebooks provided by Clayton Miller in This Repository. The results of this work has been published in their recent paper and his PhD thesis. The employed dataset is the Building data Genome Project which is an open dataset provided by their group which includes hourly whole building electrical meter data for one year from 507 non-residential buildings. You can find more detailed infomration in the provided links or the Buds lab's website