Disclosed herein is a data-driven approach for determining and presenting a more intelligent measure of the probability of failure of a substation. The disclosed approach generally involves (i) deriving respective failure probabilities of the individual assets within the substation by taking into consideration certain operating, environmental, maintenance or other types of data related to the individual assets, (ii) determining the electrical configuration of the substation, (iii) determining a substation failure probability based on the respective failure probabilities of the individual assets within the substation and the electrical configuration of the substation, and then (iv) presenting the probability of substation failure and/or the respective failure probabilities for the individual assets to a user in various ways. A user may use this probability of failure together with knowledge of the impact or consequence of a failure at the substation to make planning decisions for the substation or the electrical system.