This is far from a strict rule though, lots of shales have basically no uranium at all. Many shales come from erosion of uranium-enriched parent rock so the shale itself will have higher than typical uranium contents. Rocks like granite which are relatively rich in potassium right at formation, and rocks that come from weathering of such potassium rich rocks, tend to have elevated uranium contents, and thus tend to be associated with elevated radon in an area. Radon can become a problem when it is either generated in larger amounts in a source rock, or because the radon gets blocked after it migrates for a while. The rocks won't be highly radioactive (unlike actual uranium ore would be). The uranium content is still way lower than potassium content, though. There is a strong correlation between potassium content and uranium content in natural rocks.
(we are not talking about the fairly rare places where uranium concentrations are so high that actual uranium minerals are formed). Both are very large ions so both require a big space in the structure or they simply do not get included. Typically, the potassium spots in a structure are where the uranium ends up as a minor constituent (potassium and uranium are similar atom sizes). Uranium is a fairly uncommon element and usually occurs as a replacement element for the primary cation (positively charged element) in the mineral.
It leaves the mineral crystal and migrates upward like all low-density gases.
Radon comes from the decay of uranium, and uranium is an ion in solid minerals but radon is an inert gas so is free to move once it comes into existence (replaces the uranium parent).