Extracting Knowledge from Very Large Datasets in a Multi-Wavelength Context