An extremely useful area of statistics is a set of models that use latent variables: variables whole values we can’t measure directly, but instead have to infer from others. These latent variables can be unknown groups, unknown numerical values, or unknown patterns in trajectories.