Hadwiger's Theorem states that En-invariant convex-continuous valuations of definable sets in R n are linear combinations of intrinsic volumes. We lift this result from sets to data distributions over sets, specifically, to definable R-valued functions on R n . This generalizes intrinsic volumes to (dual pairs of) non-linear valuations on functions and provides a dual pair of Hadwiger classification theorems. n k=0 c k µ k ,The intrinsic volumes 1 µ k are characterized uniquely by (1) E n invariance, (2) normalization with respect to a closed unit ball, and (3) homogeneity: µ k (λ · A) = λ k (A) for all A ∈ S and λ ∈ R + . These measures generalize Euclidean n-dimensional volume (µ n ) and Euler characteristic (µ 0 ). This paper extends Hadwiger's Theorem to similar valuations on functions instead of sets. Section 2 gives background on the definable (o-minimal) setting that lifts Hadwiger's Theorem to tame, non-convex sets and then to constructible functions; there, we also review the convex-geometric, integral-geometric, and sheaf-theoretic approaches to Hadwiger's Theorem. In Section 4, we consider definable functions Def(R n ) as R-valued functions with tame graphs, and correspondingly define dual pairs of (typically) non-linear "integral" operators · dµ k and · dµ k mapping Def(R n ) → R as generalizations of intrinsic volumes, so that 1 A dµ k = µ k (A) = 1 A dµ k for