Perhaps it is easy to overtheorize a phenomenon such as mashups, making too much out of it, overstretching its use as metaphor. For one thing, while musical mashups sometimes can be “official” - such as when Kylie Minogue performed a mashup of her “Can’t get you out of my head” and New Order’s “Blue monday” - most of the time they are not endorsed by the artists (or the label’s lawyers), whereas astronomical data mashups are encouraged and have become routine practice. Once the principal investigators have mined the data, datasets are opened up.
Having said that, I still find the mashup music of the early 00’s - exemplified above by Dsico’s fine blend of Missy Elliott (2001) and Joy Division (1980) - something of a metaphor for what at times goes on in science. Sometimes datasets - songs - recorded for one purpose are merged with other datasets, recorded under other circumstances, to produce a work that was not thought of by the makers of the original datasets, something that builds something new and strong upon two rather disjointed things. Trivial, I know, but still I find it an interesting thing to consider; astronomers’ daily practice is sometimes about doing such mashups. Astronomy has developed a culture of making whole datasets publically available, and one part of astronomical practice is putting such stuff together.
This is not something new, something that goes on in our e-science epoch using the web as infrastructure. Take the Lund Observatory Milky Way Panorama (1955), for example. It was produced from photographs - the Franklin-Adams Atlas (1914), for example -, data for the positions and magnitudes of 7000 stars, photometric measurements of the Milky Way, data for converting RA and declination to galactic coordinates &c.
From Photographic Photometry of the Southern Milky Way (Amsterdam, 1949) by Pannekoek, one of the datasets used by Knut Lundmark and his group.
Putting these diverse datasets together to the Lund Observatory Panorama was a non-trivial exercise; an artist worked on the task for two years, painstakingly painting the Milky way, with a basis in the datasets, under the guidance of Knut Lundmark, whose aesthetic judgment seemed to have played an important role in the genesis of the picture.
Mashups of standard datasets are, as I mentioned, encouraged. There are also tools available, facilitating the mashups. Take SkyView, the frontend to many sky survey data from across the spectrum. At the query page you choose between datasets, from gamma rays to radio and can mix them up; you can choose different projections, plot coordinate grids &c.
Here is a mashup of M31 I just produced. Red and green are IRAS data, blue is the blue plate from the Palomar sky survey. Combining data from a photographic sky survey produced on Mount Palomar between 1948 and 1958 with the large Schmidt telescope there and originally recorded on photographic glass plates, 14 inches square (and in our age digitized with high-precision scanners) with data from a multi-national infrared telescope project, observing from a satellite in orbit around the earth in 1983, was just a matter of a few clicks.
