Reconstructions of the primordial power spectrum (PPS) of curvature perturbations from cosmic microwave background anisotropies and large-scale structure data suggest that the usually assumed power-law PPS has localised features (up to $\sim 10\%$ in amplitude), although of only marginal significance in the framework of $\Lambda$CDM cosmology. On the other hand if the underlying cosmology is assumed to be Einstein-de Sitter, larger features in the PPS (up to $\sim 20\%$) are required to accurately fit the observed acoustic peaks. Within the context of single clock inflation, we show that any given reconstruction of the PPS can be mapped on to functional parameters of the underlying effective theory of the adiabatic mode within a 2nd-order formalism, provided the best fit fractional change of the PPS, $\Delta\mathcal{P}_\mathcal{R}/\mathcal{P}_\mathcal{R}$ is such that $(\Delta\mathcal{P}_\mathcal{R}/\mathcal{P}_\mathcal{R})^3$ falls within the $1\,\sigma$ confidence interval of the reconstruction for features induced by variations of either the sound speed $c_\mathrm{s}$ or the slow-roll parameter $\epsilon$. Although there is a degeneracy amongst these functional parameters (and the models that project onto them), we can identify simple representative inflationary models that yield such features in the PPS. Thus we provide a dictionary (more accurately, a thesaurus) to go from observational data, via the reconstructed PPS, to models that reproduce them to per cent level precision.

We consider the possibility that the primordial curvature perturbation is direction-dependent. To first order this is parameterised by a quadrupolar modulation of the power spectrum and results in statistical anisotropy of the CMB, which can be quantified using `bipolar spherical harmonics'. We compute these for the Planck DR2-2015 SMICA map and estimate the noise covariance from Planck Full Focal Plane 9 simulations. A constant quadrupolar modulation is detected with 2.2 sigma significance, dropping to 2 sigma when the primordial power is assumed to scale with wave number k as a power law. Going beyond previous work we now allow the spectrum to have arbitrary scale-dependence. Our non-parametric reconstruction then suggests several spectral features, the most prominent at k ~ 0.006/Mpc. When a constant quadrupolar modulation is fitted to data in the range 0.005 < k Mpc < 0.008, its preferred directions are found to be related to the cosmic hemispherical asymmetry and the CMB dipole. To determine the significance we apply two test statistics to our reconstructions of the quadrupolar modulation from data, against reconstructions of realisations of noise only. With a test statistic sensitive only to the amplitude of the modulation, the reconstructions from the multipole range 30 < l < 1200 are unusual with 2.1 sigma significance. With the second test statistic, sensitive also to the direction, the significance rises to 6.9 sigma. Our approach is easily generalised to include other data sets such as polarisation, large-scale structure and forthcoming 21-cm line observations which will enable these anomalies to be investigated further.

We reconstruct the power spectrum of primordial curvature perturbations by applying a well-validated non-parametric technique employing Tikhonov regularisation to the first data release from the Planck satellite. To improve the reconstruction on small spatial scales we include data from the ground-based ACT and SPT experiments, the WiggleZ galaxy redshift survey, the CFHTLenS tomographic weak lensing survey, and spectral analysis of the Lyman-alpha forest. The reconstructed scalar spectrum (assuming the standard LCDM cosmology) is not scale-free but has an infrared cutoff at k < 5 x 10^-4 Mpc^-1 and several ~2-3 sigma features, of which two at wavenumber k/Mpc^-1 ~ 0.0018 and 0.057 had been seen already in WMAP data. A higher significance feature at k ~ 0.12 Mpc^-1 is indicated by Planck data, but may be sensitive to the systematic uncertainty around multipole l ~ 1800 in the 217x217 GHz cross-spectrum. In any case accounting for the 'look elsewhere' effect drops its global significance to ~2 sigma.

Detailed knowledge of the primordial power spectrum of curvature perturbations is essential both in order to elucidate the physical mechanism (`inflation') which generated it, and for estimating the cosmological parameters from observations of the cosmic microwave background and large-scale structure. Hence it ought to be extracted from such data in a model-independent manner, however this is difficult because relevant cosmological observables are given by a convolution of the primordial perturbations with some smoothing kernel which depends on both the assumed world model and the matter content of the universe. Moreover the deconvolution problem is ill-conditioned so a regularisation scheme must be employed to control error propagation. We demonstrate that `Tikhonov regularisation' can robustly reconstruct the primordial spectrum from multiple cosmological data sets, a significant advantage being that both its uncertainty and resolution are then quantified. Using Monte Carlo simulations we investigate several regularisation parameter selection methods and find that generalised cross-validation and Mallow's $C_p$ method give optimal results. We apply our inversion procedure to data from the Wilkinson Microwave Anisotropy Probe, other ground-based small angular scale CMB experiments, and the Sloan Digital Sky Survey. The reconstructed spectrum (assuming the standard $\Lambda$CDM cosmology) is \emph{not} scale-free but has an infrared cutoff at $k \lesssim 5 \times 10^{-4}\; \mathrm{Mpc}^{-1}$ (due to the anomalously low CMB quadrupole) and several features with $\sim 2 \sigma$ significance at $k/\mathrm{Mpc}^{-1} \sim$ 0.0013--0.0025, 0.0362--0.0402 and 0.051--0.056, reflecting the `WMAP glitches'. To test whether these are indeed real will require more accurate data, such as from the Planck satellite and new ground-based experiments.

Measurements of the SNe Ia Hubble diagram which suggest that the universe is accelerating due to the effect of dark energy may be biased because we are located in a 200-300 Mpc underdense "void" which is expanding 20-30% faster than the average rate. With the smaller global Hubble parameter, the WMAP-5 data on cosmic microwave background anisotropies can be fitted without requiring dark energy if there is some excess power in the spectrum of primordial perturbations on 100 Mpc scales. The SDSS data on galaxy clustering can also be fitted if there is a small component of hot dark matter in the form of 0.5 eV mass neutrinos. We show however that if the primordial fluctuations are gaussian, the expected variance of the Hubble parameter and the matter density are far too small to allow such a large local void. Nevertheless many such large voids have been identified in the SDSS LRG survey in a search for the late-ISW effect due to dark energy. The observed CMB temperature decrements imply that they are nearly empty, thus these real voids too are in gross conflict with the concordance LCDM model. The recently observed high peculiar velocity flow presents another challenge for the model. Therefore whether a large local void actually exists must be tested through observations and cannot be dismissed a priori.

Detailed analyses of the WMAP data indicate possible oscillatory features in the primordial curvature perturbation, which moreover appears to be suppressed beyond the present Hubble radius. Such deviations from the usual inflationary expectation of an approximately Harrison-Zeldovich spectrum are expected in the supergravity-based 'multiple inflation' model wherein phase transitions during inflation induce sudden changes in the mass of the inflaton, thus interrupting its slow-roll. In a previous paper we calculated the resulting curvature perturbation and showed how the oscillations arise. Here we perform a Markov Chain Monte Carlo fitting exercise using the 3-year WMAP data to determine how the fitted cosmological parameters vary when such a primordial spectrum is used as an input, rather than the usually assumed power-law spectrum. The 'concordance' LCDM model is still a good fit when there is just a 'step' in the spectrum. However if there is a 'bump' in the spectrum (due e.g. to two phase transitions in rapid succession), the precision CMB data can be well-fitted by a flat Einstein-de Sitter cosmology without dark energy. This however requires the Hubble constant to be h ~ 0.44 which is lower than the locally measured value. To fit the SDSS data on the power spectrum of galaxy clustering requires a ~10% component of hot dark matter, as would naturally be provided by 3 species of neutrinos of mass ~0.5 eV. This CHDM model cannot however fit the position of the baryon acoustic peak in the LRG redshift two-point correlation function. It may be possible to overcome these difficulties in an inhomogeneous Lemaitre-Tolman-Bondi cosmological model with a local void, which can potentially also account for the SN Ia Hubble diagram without invoking cosmic acceleration.

Observations of anisotropies in the cosmic microwave background by the Wilkinson Microwave Anisotropy Probe suggest the possibility of oscillations in the primordial density perturbation. Such deviations from the usually assumed scale-free spectrum were predicted in the multiple inflation model wherein `flat direction' fields undergo rapid phase transitions due to the breaking of supersymmetry by the large vacuum energy driving inflation. This causes sudden changes in the mass of the (gravitationally coupled) inflaton and interrupts its slow roll. We calculate analytically the resulting modifications to the density perturbation and demonstrate how the oscillations arise.