Why is the mapped universe shaped like an hourglass? I've watched a video from the American National History Museum entitled The Known Universe.
The video shows a continuous animation zooming out from earth to the entire known universe. It claims to accurately display every star and galaxy mapped so far. At one point in this video [3:00 - 3:15 minutes] it displays this text:

The galaxies we have mapped so far.
The empty areas where we have yet to
  map.

At this point, the shape of the "universe" is roughly hourglass, with earth at it's centre.

I'm having trouble understanding what this represents and why they shape is hourglass. Is this simply because we have chosen to map in those directions first? Is there a reason astronomers would choose to map in this pattern, or is this something more fundamental about the shape of the universe? Or is it to do with the speed of light reaching us from these distant galaxies?
Continuing on from the hourglass pattern, the cosmic microwave background radiation is represented as a sphere and labelled "Our cosmic horizon in space and time". This doesn't help clear anything up. If we can map CMB in all directions, why have we only mapped galaxies in this hourglass shape.

 A: Someone will correct me if I'm wrong, but I think the explanation is the following:
Our planet is part of the Solar system, which is itself rotating around the center of the Milky Way in one of its spiral arms. Whenever we want to look at deep space in the direction of these arms, we have to look through thick layers of dust and other stars from the different spiral arms our line of sight will cross. So basically, our own galaxy is impeding our view in certain directions. So the cones you see are basically the up and down direction with respect to the galactic plane.
A: The correct answer is actually given by the person who asked the question (as well as Scott Wales), while most of the answers above give a misleading one. The shape is hourglass simply because Sloan Digital Sky Survey and 2 degree Field Sky Survey did not map out the whole sky, but only about 1/4 of it (for the SDSS). The other directions therefore appear dark in the visualization. 
There is of course nothing special about the directions in which these surveys look. In fact, there exist other galaxy surveys that cover nearly the whole sky, though I guess they haven't produced fancy visualizations. (Moreover different surveys operate at different wavelengths of light and are complementary in this, and several other, ways). 
It is true that the Milky Way would obstruct some portion of the sky, but 1) this is a small fraction (20% or so of the sky), and 2) if you showed the sky in the Milky Way direction, it would actually appear bright, not black, for obvious reasons.
A: The visualisation shows an 'hour glass' shape because you're looking at it from above. The actual layout of the points is like two pie pieces put point to point. The visualisation is of the distribution of galaxies in the universe at large distances, and has either come from the Sloan digital sky survey (SDSS) or the 2 degree field galaxy redshift survey (2dfGRS).
Making these plots requires a large amount of telescope time, as for each galaxy you must measure the spectra of the light it emits accurately enough to get the redshift, which is a measure of how distant it is. The surveys thus don't cover the whole sky, instead they are done as stripes, which leads to the pie shape when projected through the universe. In order to observe the structure that forms from the galaxy distribution you want a continuous sample which goes over a large angular range, which the stripe gives in one direction. You also want to make sure that you aren't sampling a special region of sky, hence there's another sample taken on the other side of the sky.
In short, the shape of the sample is purely due to the survey format, the actual structure of galaxies will fill a sphere. 
A: The intuitive reason why there is a black strip in the area corresponding to where Milky Way lies is explained in the other answers but this is far from being the end of story.
In particular, why doesn't it show also on WMAP picture? Well, the short answer is it does. But what they do (and what hasn't been answered in other questions) is that they scan different wavelengths of light and that gives them enough information (by using various spectral methods) to subtract the effects of Milky Way (and also other objects that obstruct the view).
So it's obvious that one can't really say that we don't see through Milky Way. We don't see only in the narrow sense of "see" as in seeing in visible spectrum of light but obviously this is not physical at all. Visible light isn't any special just because we, humans, happen to have eyes that register it. Astronomers observe the universe at very different wavelengths (like X-ray radiation, gamma bursts, IR, UV and others). And even this would be too narrow. They observe all kind of particles other then light. Neutrinos are one particular type that is obviously not very much bothered by our galaxy.
A: As it has been pointed out, the Milky Way is to blame for the hourglass shape of the maps of the universe. I would like though to point out for the case of the CMB and the WMAP images that have the contribution of the Milky Way removed, that there seem to be some open issues. There is a very nice lecture given at PI on the subject of CMB anomalies inferred from WMAP, that discuses the possibility that they could have something to do with the way that the signal from the Milky Way is subtracted. The lecture is:
If the CMB is right, it is inconsistent with standard inflationary Lambda CDM (2008, Glenn Starkman) 
and a relevant paper:
Large-angle anomalies in the CMB
A: First of all, the universe is most certainly not shaped like an hourglass. It simply looks that way because the gas and dust in the plane of our galaxy obstruct our view of anything outside the galaxy in those directions. So we can only see other galaxies (and similarly distant objects) by pointing telescopes at some angle to the galactic plane. That gives the "hourglass" shape: it's simply because those are the only directions we can see in. In reality, we have every reason to think galaxies are distributed more or less uniformly, once you look at a large enough scale.
The video description doesn't cite its sources, but I suspect that (some/most of) the information on the distant galaxies comes from the Sloan Digital Sky Survey, which is AFAIK the most comprehensive survey of objects in the universe outside our own cluster of galaxies. You might want to check out the information on their website if you're interested in this stuff.
And as long as I'm citing sources, the latest CMB data comes from the WMAP project.
