I think it's just a math artifact. There's lots of "holes" in the orbit inclinations -- their distribution is actually very non-uniform. But at a frozen moment in time, the object's latitude will be anywhere from -i to +i, so the latitude distribution is "smeared out" over the whole range. The gap in the polar inclinations is the only one that's not, so that's the latitude gap you see.
Here's a histogram of the orbit inclinations where you can see this:
Right - even though polar orbits are pretty useful, so there probably are quite a few satellites with inclinations close to 90 degrees, they spend less than 10/360 = 2.8% of their time within 5 degrees of the north pole. What's surprising is the sudden increase in the density of stuff around 80 degrees north.
One possible explanation is the combination of the iridium 33 collision debris field, which is in orbits with inclinations around 86 degrees, and Fengyun 1C debris, which is around 99 degrees (so, 81 degrees but going the opposite direction). The limit of the distributions of those two clusters might account for the clear 'edge' of the circle around the pole.
Here's a histogram of the orbit inclinations where you can see this:
https://i.imgur.com/K9zsCy3.png
Here's the raw orbits data that the site uses:
http://stuffin.space/TLE.json
https://en.wikipedia.org/wiki/Two-line_element_set