REVIVAL

Recovery of Missing Values in Time Series

Summary

The REVIVAL project focuses on developing methods that benefit researchers conducting data analyses requiring complete time series datasets. In fields such as hydrology, sensors are deployed to monitor environmental variables, temperature, air pressure, precipitation, wind speed, and humidity over time. However, these sensors often experience temporary failures, resulting in datasets with missing data blocks. Since hydrological models used for weather forecasting and environmental predictions require complete datasets, hydrologists must first reconstruct these missing values before proceeding with their analyses.

Another major challenge in hydrology is scalability. Managing large volumes of data, through processes like backup, import, export, or copying is complex and resource-intensive. As a result, hydrologists often work with truncated datasets, limiting the scope and accuracy of their models. The REVIVAL project addresses this issue by providing techniques to recover data that would otherwise be excluded, thereby enabling the use of complete datasets. Effective data recovery will also enhance the performance and interactivity of hydrological systems such as flood warning and drought monitoring tools.

Grant Details

REVIVAL is a 50’000 CHF one-year project funded by the Hasler Foundation. The project was led by Mourad Khayati as the main investigator and sole recipient.