Access is obtained to raw historic information technology migration data, which data is transformed in accordance with a classification scheme to obtain classified key historic information technology migration data. A statistical model with conditional distributions of the key historic information technology migration data is built. Prior to a new information technology migration project, access is obtained to incomplete data characterizing the new information technology migration project. The statistical model is used to derive data missing from the incomplete data characterizing the new information technology migration project, to obtain derived data. At least one of a migration plan and an estimate for the new information technology migration project is created based on the incomplete data and the derived data.