In this paper, land prices for nonresidential real estate listings in Windsor, Ontario, Canada, are analyzed with the use of ordinary least squares regression models. Location and transportation attributes are used in the analysis to explain observed land prices. Seven models were estimated to control the heterogeneity in the land use types. The results show differences in factors explaining land prices of the models, indicating that caution should be used when land use types are aggregated together. The role of transportation carried mixed results. Rail had a positive effect for industrial properties but was negative for commercial and food services. Transit had a positive effect on vacant land. Direct proximity to highway ramps had a negative correlation for vacant land, but indirect variables such as potential (residential) accessibility and time to the central business district indicated the positive effects that roads and highways had. Testing showed that although spatial autocorrelation was present in the price data, the independent regressors used in the modeling partially mitigated that effect. Multicollinearity and heteroscedasticity were also accounted for throughout the modeling process. The results obtained provide a useful account of various spatial and transportation-related phenomena for a midsize Canadian metropolitan area. Moreover, the empirical analysis is particularly valuable given the lack of modeling done on the commercial and industrial prices when compared with that done for residential properties.