Query driven visualization of large scale multi-dimensional astronomical catalogs

PhD ceremony: Mr. H. Buddelmeijer, 13.15 uur, Doopsgezinde kerk, Oude Boteringestraat 33, Groningen

Dissertation: Query driven visualization of large scale multi-dimensional astronomical catalogs

Promotor(s): prof. E.A. Valentijn, prof. S.C. Trager, prof. J.B.T.M. Roerdink

Faculty: Mathematics and Natural Sciences

Billions of sources are detected with modern telescopes, for which hundreds of parameters are quantified. New techniques are necessary for interactive exploration of these large datasets. A query driven way of visualization is introduced in this thesis to achieve both scalability and flexibility.

An important aspect in this is storing information on how the astronomical data is processed. Not only does this clarify how existing datasets have been derived, but it also allows new datasets to be generated automatically. The result of this is that discovering existing data and creating new data is performed in the exact same manner: by requesting it. Data requests can be formulated independent of implementation. This makes it possible to request the data from within the visualization software itself.

New data is only created when the existing data is not suitable to fulfil such a request. This new data will automatically be defined such that it is most suitable for reuse for later requests. However, only the part of the data that is required for the visualization is actually processed, ensuring scalability. Furthermore, data is only stored if reprocessing on demand is too time-consuming.

Neither the scientist nor the visualization software has to worry about creating data in a scalable and reusable way, because this is all automated. Therefore, query driven visualization gives scientists more time for the research they want to perform with their data, because the data administration is implicit.