Determining transport coefficients for galactic cosmic ray (GCR) propagation in the turbulent interplanetary magnetic field (IMF) poses a fundamental challenge in modeling cosmic ray modulation processes. GCR scattering in the solar wind involves wave‐particle interaction, the waves being Alfven waves which propagate along the ambient field (B). Empirical values at 1 AU are determined for the components of the diffusion tensor for GCR propagation in the heliosphere using neutron monitor (NM) data. At high rigidities, particle density gradients and mean free paths at 1 AU in B can only be computed from the solar diurnal anisotropy (SDA) represented by a vector A (components Ar, Aϕ, and Aθ) in a heliospherical polar coordinate system. Long‐term changes in SDA components of NMs (with long track record and the median rigidity of response Rm ~ 20 GV) are used to compute yearly values of the transport coefficients for 1963–2013. We confirm the previously reported result that the product of the parallel (to B) mean free path (λ||) and radial density gradient (Gr) computed from NM data exhibits a weak Schwabe cycle (11y) but strong Hale magnetic cycle (22y) dependence. Its value is most depressed in solar activity minima for positive (p) polarity intervals (solar magnetic field in the Northern Hemisphere points outward from the Sun) when GCRs drift from the polar regions toward the helioequatorial plane and out along the heliospheric current sheet (HCS), setting up a symmetric gradient Gθs pointing away from HCS. Gr drives all SDA components and λ|| Gr contributes to the diffusive component (Ad) of the ecliptic plane anisotropy (A). GCR transport is commonly discussed in terms of an isotropic hard sphere scattering (also known as billiard‐ball scattering) in the solar wind plasma. We use it with a flat HCS model and the Ahluwalia‐Dorman master equations to compute the coefficients α (=λ⊥/λ∥) and ωτ (a measure of turbulence in the solar wind) and transport parameters λ||, λ⊥, Gr, Gθs, and an asymmetric gradient Gθa normal to the ecliptic plane. We study their dependence on rigidity (R), p/n intervals, sunspot numbers (SSNs), and solar wind parameters at 1 AU. λ|| exhibits a strong 22y dependence but Gr does not, explaining solar polarity dependence of λ|| Gr. The computed Gr values are an order of magnitude greater than those reported by our colleagues making an ad hoc assumption that α is low (0.01). At high rigidities, the drift contribution at 1 AU is small and unsteady. A new methodology is outlined to compute yearly GCR north‐south anisotropy (Aθ) from the data for a single detector sorted for p/n intervals. We show that Gθa is the main contributor to Aθ in the steady state, and Gθa is shown not correlated with the north‐south excess SSNs.