We propose nonparametric kernel estimation procedures for panel data to investigate heterogeneous dynamics across individuals. We first estimate the sample mean, autocovariances, and autocorrelations for each individual, and we then consider kernel density, cumulative distribution function, and quantile estimators based on them. We investigate asymptotic properties of the proposed estimators, and propose a bias correction method and a bootstrap inference procedure. Monte Carlo simulations and an empirical application illustrate the usefulness of the proposed procedures.