Public transport punctuality affects people’s experiences of public transport and hence how they choose to travel. Service reliability is typically poorly accounted for in computational analyses that quantify expected travel time. By constructing models that do not consider scheduling or the consequences on passenger waiting time of services being early or late, public transport forecasting and accessibility models often abstract how passengers really experience public transport. This paper presents the progress made on a current Master’s thesis in Geographic Information Systems at Victoria University of Wellington. Combining standardised transit data with a large database of real-time information and Snapper smartcard passenger transactions, transit models of Wellington are built using open source tools. These models will be able to include the consequences of observed system unreliability in the quantification of expected passenger waiting and travel time. I discuss the mechanisms, opportunities and limitations of this approach, including the development of the models. Hypothetical accessibility analyses constructed with and without accounting for unreliability will be used to determine whether including system unreliability leads to significantly different results in accessibility analyses and hence whether future accessibility analyses should include it.