My next assignment for Insight is to create tables that will be used for a regression analysis to predict values based on various property features. The blueprint for this analysis is found in the scholarly article [first picture].
Basically, there is a data set of comparable sales of industrial properties that each have their own unique characteristics (land size, building size, percent office finish, clear height in the warehouse, etc.). The first table summarizes the data set, and the subsequent tables analyze each of the property features so that each comparable can be ranked (for each feature) [second picture]. Then a weighted average of the features is calculated. Once that is done, we’re going to use the Solver add-in in Excel, which will determine the optimal weights for the various features so that the variance of the weighted averages is minimized [last picture]. This results in an objective weighting of the features, because math.
The real benefit is that it’ll provide calculated support for the appraiser’s conclusions, rather than just relying on the appraiser’s opinion of how each feature contributes to the value.
