This talk will be a combination of 4 papers. I will start from the best-studied univariate mean change point detection problem [1], which serves as a blueprint for more complex settings, and then talk about high-dimensional covariance [2], sparse dynamic networks [3] and distribution functions (ongoing project) change point detection problems. I will show the phase transitions, minimax optimality and binary segmentation type algorithms in all settings. To conclude the talk, I will talk about the penalisation techniques used in the change point detection problem and the limitations thereof.