# Form of citation for data via Wolfram Mathematica

What citation or reference should I adhere to if I'm using curated data obtained via Mathematica? I'm talking about data, or eg. plots thereof, obtained using functions such as SatelliteData, PlanetData, etc.

I am aware that for:

But this doesn't specifically cover the blindspot described above.

I'm interested in using the data, or calculations and visualizations made from the data, for two types of communication:

• Use in presentations for education (eg. a plot of some orbit, a table of planet masses)
• use in research aimed at published journals
• For the former AstronomicalData[], there was this list of references; now that it's been split up into smaller curated data functions, I'm not sure. I think it's still best to just use data from actually citable sources, and if they happen to agree with what Mathematica spits out, then note that observation. – J. M.'s ennui Aug 23 '15 at 15:16
• It would be desirable to have an answer from somebody at WRI! – murray Aug 24 '15 at 0:11
• Thanks @J.M. For publication or serious computation, using actual citable sources does appear to be the only way at the moment. However, it seems to negate the whole point of "curated" data if you can't be certain of its provenance. If I knew that I could just cite "Wolfram Research" or "Wolfram|Alpha", that would already be a step forward, certainly for presentations / education. – TransferOrbit Aug 26 '15 at 6:46
• I see "curated data" at the moment as a convenient way to try out any computations you have with "real world data" without having to go through the trouble of searching for sources. But, as soon as you're no longer doing "toy computations" I think it behooves the researcher to actually have to do the legwork for reliable data sources. – J. M.'s ennui Aug 26 '15 at 6:49

Wolfram has already got you covered actually. Check out the choices for annotation commands you can give to Mathematica with PlanetData of SatelliteData from the documentation, with particular focus on "Source":

PlanetData["Earth", "Radius"]


• Thank you, I wasn't aware of this. I'm afraid it still leaves a lot of questions open for me. Just for this query, there appear to be two sources. If I were to cite the use of this Radius, then I still don't know which source this comes from. In addition, I saw somewhere that Wolfram sometimes modifies the data in some way, so the definition of "curating" is also open. Given all this opacity, it would therefore seem to make more sense to cite "Wolfram Research", or something similar, but it's not clear if that's correct, or "Wolfram|Alpha", or something else. – TransferOrbit Aug 26 '15 at 6:37
• @zentient Yeah, figuring out to what extent the curation happens is indeed a major pain in the ass (I'm guessing they mostly just do some sort of standardisation for all the values from different, but that is just guessing). If I were to quote the data, I would probably list Mathematica and the two sources as the sources for the value, or, it producing only one source, I would find the same data from Wolfram|Alpha and provide the according citation (as in the link you provided; thanks for that by the way). – V-J Aug 26 '15 at 7:32
• And if you had to provide a citation for further processed data (plots, etc.), in that case it would maybe be the simplest to list Mathematica and each of the sources (despite a single plot probably producing the mother of all citation lists)? – V-J Aug 26 '15 at 7:36
• @zentient One problem with citing W|A is that it is in continuous flux. A couple of years later it might give different (possibly updated) answers. And the old versions are no longer accessible. It's not the same as citing "Table of Quantities, 3rd edition, 1987", which can always be checked out of the library. Another problem is of course that it is not an original source, yet it provides no practical way to trace the data to the original source. These things are very frustrating. But do you see a way to have the convenience of W|A without all these problems? I don't really see one ... – Szabolcs Oct 22 '15 at 10:46
• @Szabolcs I guess the best way to go with W|A would be to, obviously, note the date when it was retrieved, but that would still leave the issue with accessing old data. I'd say the way to go would be to use Mathematica over W|A for the (more) accurate source, provide the source with date, and theoretically the person checking it back could try to find the latest publication of the source going back to the date when the value was retrieved. Not as easy as directly using the source publication, but in the end we would just be debating which outdated source do we want to double check. – V-J Oct 22 '15 at 10:55