Presence of islet autoantibodies precedes the onset of type 1 diabetes but it does not predict whether and how fast symptomatic disease appears. Here authors present a model to predict and visualize progression to diabetes by using a large longitudinal data set on autoantibodies and clinical parameters as input.