Localization generalizes the construction of fractions. Given a commutative ring R and a multiplicatively closed subset S, the localization S⁻¹R consists of formal fractions r/s with r ∈ R, s ∈ S, subject to the same equivalence relation as ordinary fractions. This construction lets you "zoom in" on the behavior of a ring near a prime ideal by inverting everything outside it, reducing global questions to local ones.
The construction of the rational numbers from the integers — forming fractions a/b with b ≠ 0 — is the prototype of localization. In commutative algebra, localization generalizes this by letting you choose which denominators to allow. Given a commutative ring R and a multiplicatively closed set S ⊆ R (meaning 1 ∈ S and if s, t ∈ S then st ∈ S), the localization S⁻¹R consists of formal fractions r/s, where two fractions r/s and r'/s' are identified if there exists u ∈ S with u(rs' - r's) = 0.
The two most important cases are localization at a single element and localization at a prime ideal. Localizing at an element f means taking S = {1, f, f², ...}, producing R_f = R[1/f], the ring where f becomes invertible. Localizing at a prime ideal 𝔭 means taking S = R \ 𝔭, inverting everything outside 𝔭. The result, written R_𝔭, is a local ring — a ring with exactly one maximal ideal. This is the most powerful application: it lets you study the behavior of R "near 𝔭" by making all other prime structure invisible.
Localization has excellent algebraic properties. It is an exact functor: it preserves short exact sequences of modules, meaning it commutes with taking kernels, images, and cokernels. It also commutes with taking quotients, sums, and intersections of ideals. The ideal structure of S⁻¹R is simpler than that of R: the prime ideals of S⁻¹R correspond exactly to the prime ideals of R that are disjoint from S. When localizing at 𝔭, this means the primes of R_𝔭 are exactly the primes of R contained in 𝔭, with 𝔭 itself becoming the unique maximal ideal.
The local-global principle is the philosophical payoff. Many properties of a ring or module hold globally (over R) if and only if they hold locally (over R_𝔭 for every prime 𝔭). For instance, a module is zero if and only if it is zero after localizing at every prime. An R-module homomorphism is injective (or surjective) if and only if it is so after every localization. This reduces hard global questions to easier local ones, where you work in a ring with a single maximal ideal and can exploit the special structure of local rings.