Airport sponsors, typically municipal governments in the US, along with the Federal Aviation Administration (FAA) engage in a number of planning activities to determine the long-term development needs of airport infrastructure. One of the primary tasks of these airport planning activities is to estimate future use of the airport. Airport planners use two broad categories of methods to estimate future use of the airport; 1) peer group learning (as in, considering the experiences of “peer” airports) and 2) aviation demand forecasting. Airport master planning, a federally mandated planning process for airports for infrastructure planning such as building a new runway, for instance, relies on these techniques to be effective. Yet, there are numerous challenges to how airport planners can use these techniques effectively. These challenges can be largely categorized as the problem of demand uncertainty and optimism bias; demand uncertainty stemming from the dynamic socioeconomic and aviation industry trends and optimism bias from the economic development narrative surrounding airports and the federal funding incentives for airport infrastructure projects. Demand uncertainty and optimism bias create large forecast errors and have led airport planners to make unwise infrastructure investment decisions. In this dissertation, I use publicly available aviation and census data to develop and test new methodologies that enable airport planners to 1) identify true airport peers that share similar socioeconomic trends, 2) predict the probability of a severe contraction in passenger volumes in the next 10 years, and 3) improve forecast accuracy by incorporating past forecast errors systematically into the current forecast and “ground” optimistic forecasts. I show that the methodologies can have much more immediate and robust impact on airport planning than traditional methods to curtail demand uncertainty and optimism bias.