If you want to get the "wow" factor from a science project.... use a gui of some sort. Of course I recommend Tk or Gtk2. ;-) How about something cool like simulating an instrument on a satellite and logging the data on a strip chart recorder on earth? See Tk Realtime data aquisition Maybe monitor solar wind? You could connect right to the spacecraft data-servers like solar data server Have the data recorder and maybe make a 2-d visualization of the magnetosphere at the current moment, that could be done with animated gifs.

You can get most of that data in closer to real time from CDAWeb, and it'll generate plots, too. (I'm not a fan of their interface, though ... you can also get their data via a web service or see VSPO for other sources. I know of other solar weather plotting software ... but much of it's not in Perl (IDL, Java, etc) or is tied up with legal questions about distribution.

One that is more real-life, and could lead to a career ...

I'm just going to assume that you're just saying that heliophysics data won't lead to a career as they normally have physics phds write the code ... who have no formal programming education. (and they tend to write data systems that only answer their own questions, without considering data reuse ... but I'll probably complain about that more at the fall AGU)

... and as my boss just saw me looking at NASA websites and asking what I was doing, he suggested taking a look at the Stanford Solar Center, as they've had students turn some of their project ideas into winning science fair projects.