Alignments / Crosswalks: How
Guide to Creating Spreadsheets for Crosswalking Datasets
1. Decide which datasets are being crosswalked.
2. Decide which dataset is going to be the root and which is going to be the target. If you are going to be crosswalking one dataset with several others, then it makes sense to make the former the root and the latter ones the target.
3. Create a spreadsheet with column headings indicating which columns are the Classes and Properties of the root dataset and of the target dataset. If the root and/or the target dataset has more than two levels, then make sure to include them and label them appropriately.
After the root and target column headings have been created, create a column with the heading Mapping. This column will explain the relationship between each root and target. The following are the relationships that will be used in the mapping area:
Equivalent: This indicates that the root element and target element are the same.
Close Match: This indicates that the root and target elements are a close match to one another. This usually occurs when the two elements are from different vocabularies, but they mean or refer to the same thing (e.g. foaf:depiction and mo:image are a close match because they both refer to a picture of something).
Related: The indicates that while the root element and target element are related. This usually means that they are the same kind of thing just different examples of it (i.e. both elements are websites).
Narrower: This indicates that the root element covers an aspect in which the target element is a part of (e.g. foaf:page is broader than foaf:homepage because foaf: homepage is a specific type of foaf:page, hence narrower than foaf:page).
Broader: This indicates that the root element is part of an aspect which the target element covers.
Inverse: This indicates that the root element and the target element have an inverse relationship (e.g. foaf:primaryTopic is an inverse of foaf:isPrimaryTopicOf.
Special: This indicates that there is some kind of relationship between the root and the target that is not covered by the other Mapping choices. The extent of the relationship should be noted in the Notes column.
The column after the Mapping column should be named Notes. This is for any information about the two elements compared that is not covered in the rest of the columns.
4. Type the root Classes and Properties into the appropriate columns or copy and paste them from an existing spreadsheet.
5. Match the target Classes and Properties with those that have a relationship with a root Class or Property. It is recommended that you look for equivalent Classes and then look for any of their Properties that have a common relationship. If you find any of these, put the equivalent Classes on the same row and any Properties that have a relationship on the same row. After doing this for all the Properties for equivalent Classes, match all the other Properties even if they do not have a common class. When you do this, match the remaining target Properties to ones that they have a relationship with even if they do not have a class in common. When you do this, put the Class of the target property in the Target Class column on the same row and in brackets.
This work was supported by a grant from IMLS. It is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License. reserved.