Sparse matrices are rarely discussed, if at all, due to the relative complexity of implementation and maintenance; however, they may be a very convenient and useful instrument in certain cases. Basically, sparse matrices are conceptually very close to arrays, but they're much more efficient when working with sparse data as they allow memory savings, which in turn allows the processing of much larger amounts of data.
Let's take astrophotography as an example. For those of us not familiar with the subject, amateur astrophotography means plugging your digital camera into a telescope, selecting a region in the night sky, and taking pictures. However, since pictures are taken at night time without a flashlight or any other aid (it would be silly to try to light celestial objects with a flashlight anyway), one has to take dozens of pictures of the same object...