Explore chapters and articles related to this topic
Applying human mobility and water consumption data for short-term water demand forecasting using classical and machine learning models
Published in Urban Water Journal, 2020
Kamil Smolak, Barbara Kasieczka, Wieslaw Fialkiewicz, Witold Rohm, Katarzyna Siła-Nowicka, Katarzyna Kopańczyk
To ensure maximum efficiency in data processing, a spatiotemporal database was created using the open-source database software PostgreSQL and its extensions: PostGIS for spatial data and TimescaleDB for time-series. All the data have to be transformed and loaded into a database predefined structure. This has to be done manually. Therefore, the model is independent of the structure of provided data. Furthermore, the data are aggregated spatially into DMAs and temporally to one-hour bins. If provided data (either water consumption or mobility data) have lower temporal resolution it can be aggregated to the larger bins. Similarly, if an infrastructure manager provides different spatial segmentation (i.e. DMAs), then the data are aggregated to the provided areas. The datasets used in this study were thoroughly tested for their reliability, representativeness and consistency.