A population of N neurons encodes stimuli on a low-dimensional manifold. PCA reveals the intrinsic dimensionality. Tune the latent dimension, noise, and tuning curves to see how the neural code geometry changes.