Looking for Mollweide projection astronomy software

[m74z00219]

Hi all,

I would like to make a pretty Mollweide projection all-sky plot, but I'm having trouble finding software. I would like to feed in J2000 equatorial coordinates in sexagesimal representation, spectral type, and distance into some piece of software that could return the Mollweide projection plot with the spectral type indicated by a color code and the distance indicated by bubble size.

I will settle for any piece of software that can handle at least some of my desires. Minimal coding is preferred. No coding required would be ideal. If there is some online tool that I could feed data and spit out a plot for me, that would be nice. If coding is needed, I can do matlab.

Basically, I want a nice "cherry" to put on top of my thesis.

Any suggestions?


Thanks,
M74

[grapes]

Just out of curiosity, why Mollweide?

[m74z00219]

Just out of curiosity, why Mollweide?

The various data I'm using derives from observations scattered across the whole sky, so I wanna make sure not to give preference to any particular area of the sky. Plus, I think they look pretty


M74

[ngc3314]

If someone around there has an IDL license, one of its projection options is Mollweide, and a routine with loop can make individual symbol sizes and colors for each point.

[Hornblower]

The various data I'm using derives from observations scattered across the whole sky, so I wanna make sure not to give preference to any particular area of the sky. Plus, I think they look pretty


M74

What do you mean by "giving preference?"

[grapes]

What do you mean by "giving preference?"

Probably, that Mollweide is an equal area projection.

There are other equal area projections, that are not as complicated to compute (a non-issue, if you let the computer do it for you :) ). So, is it that the data is pretty, or the Mollweide is pretty? :)

[m74z00219]

What do you mean by "giving preference?"

What grapes said.

There are other equal area projections, that are not as complicated to compute (a non-issue, if you let the computer do it for you :) ). So, is it that the data is pretty, or the Mollweide is pretty? :)

Oh, if there are other equal area projections, I am very much open. It's looking like I'm going to have to write a little matlab script at this point anyway. What do you suggest? The data is interesting enough, I suppose, but this allsky map is meant to be more of a nice aesthetic. However, I am a firm believer in the power of communicating data well.

I highly recommend Rosling's TED talk: http://www.youtube.com/watch?v=hVimVzgtD6w


M74

[Jeff Root]

I'm pretty sure the COBE and WMAP maps used Mollweide
projection. Unless you find something that seems distinctly
better and want to champion it, I'd stick with what the other
guys are using.

-- Jeff, in Minneapolis

[grapes]

Oh, if there are other equal area projections, I am very much open. It's looking like I'm going to have to write a little matlab script at this point anyway. What do you suggest?

Probably the Lambert cylindrical. All you do is take the sine of each latitude and plot--but, as I said, if you have to program this, any projection will be about the same effort. I would *not* use the Peters, a rough comparison of it to a globe is jarring (and, it's politically charged). (The Peters is a special case of the Gall, anyway--as is the Lambert cylindrical)

http://en.wikipedia.org/wiki/Map_projection#Equal-area

I'm pretty sure the COBE and WMAP maps used Mollweide
projection. Unless you find something that seems distinctly
better and want to champion it, I'd stick with what the other
guys are using.

Definitely true.

[m74z00219]

Well, I decided on something called the Hammer projection, which is like the Mollweide but with no parallels.

http://en.wikipedia.org/wiki/Hammer_projection


I hope the equations are correct.


Anyhoo, I attached a plot with fake data (random dots). Looks like this is gonna work. The real leg will be converting my sexagesimal coordinates (ra in hours, dec in degrees) into radians.


I'll post the final result when I'm done.


Thanks everyone,
M74

[grapes]

Well, I decided on something called the Hammer projection, which is like the Mollweide but with no parallels.

http://en.wikipedia.org/wiki/Hammer_projection

Nice, that's a lot more complicated than the Lambert, but of course once you start...

It is elliptical in shape, like the Mollweide, but you don't have to iterate in the calculations. Possibly, in the gross results, it'd be difficult to tell much of a difference.

[Jeff Root]

Why are the random dots bunched near the poles?
I wouldn't think that would happen in an equal-area
projection. (In a Robinson projection, I'd expect
the dots to be more sparse near the poles.)

-- Jeff, in Minneapolis

[m74z00219]

Why are the random dots bunched near the poles?
I wouldn't think that would happen in an equal-area
projection. (In a Robinson projection, I'd expect
the dots to be more sparse near the poles.)

-- Jeff, in Minneapolis

I'm not really sure what I thought "equal area" meant before, but what it actually means in this context is this: the ratio of sky area A to its projected area A' is equal to the ratio of sky area B to its projected area B'. I suppose the random dots had to appear denser at the bottom and top due to all lines of longitude meeting at the poles.

http://www.progonos.com/furuti/MapPr.../areaPres.html


As promised, here's what I have so far. This is real data. It reads left to right, right ascension (0 to 2*pi radians), and bottom to top, declination (-pi/2 to pi/2 radians).

To give a little context, the "hotter" or "whiter" the color the earlier the spectral type (M6 to Y2). Size of circles is a redundant indicator in here.

[grapes]

Anyhoo, I attached a plot with fake data (random dots). Looks like this is gonna work. The real leg will be converting my sexagesimal coordinates (ra in hours, dec in degrees) into radians.

I'm not really sure what I thought "equal area" meant before, but what it actually means in this context is this: the ratio of sky area A to its projected area A' is equal to the ratio of sky area B to its projected area B'. I suppose the random dots had to appear denser at the bottom and top due to all lines of longitude meeting at the poles.

Ah, I see. I was like Jeff--when you said fake random data, I was expecting data uniformly distributed over the sky, not uniformly distributed in lat/lon. I'd also noticed the "lack" of points along the equatorial regions.

Your final product seems much better, from that standpoint. Although, it still has clumps, and voids.

[parejkoj]

For the future, the matplotlib basemap module does all kinds of map projections, including mollweide. It was originally written to make plots of the Earth, but can easily work with astronomical data as well (e.g., our reversed longitude convention). The examples directory includes an "allskymap.py" file with some additional tricks to that end.

[m74z00219]

Your final product seems much better, from that standpoint. Although, it still has clumps, and voids.

Thanks. As for the clumps and voids...it might reflect a bias in what parts of the sky were searched. A lot of these objects are 2MASS and WISE and those are both all sky surveys. However, I also have a lot of objects from VLA, CFHT (Hawaii), and CTIO (Chile). IDK for sure, but this is something to think about.

I was also thinking that they might be loosely matching the Milky Way, but that can't be since they're all nearby objects (<25pc)

For the future, the matplotlib basemap module does all kinds of map projections, including mollweide. It was originally written to make plots of the Earth, but can easily work with astronomical data as well (e.g., our reversed longitude convention). The examples directory includes an "allskymap.py" file with some additional tricks to that end.

Thanks for the info!