Section 15 data has already proven itself a useful tool in transit decision making. Yet, its wider use has been inhibited by the difficulty of accessing it electronically. This paper describes a set of strategies for extracting, reorganizing and evaluating data originating in the electronic data files disseminated by Transportation Systems Center on magnetic tape.
The current organization of information within the files is unsuitable for most statistical software packages. Therefore, it is necessary to extract information from the Section 15 files and rearrange it in a form suitable for analysis. Different classes of missing data are also defined and remedies for the problem are addressed. Additionally, the cross-validation of values and the computation of basic transit variables are considered.
Many statistical models make assumptions about the distributional characteristics of variables. Differences of scale among transit systems on such measures as size of fleet often result in variables whose distributions violate these assumptions. Transformations which remedy the problem are recommended.