INFORMATION CONTROL
Auditing internal Revit information
Looking at another part of the project
Data exported from Navisworks were useful for reviewing element properties, but they left out part of a Revit file's internal organization. Views, worksets, templates, and families are part of how a team builds and manages its models; they do not necessarily reach a federation in the same level of detail. I began developing an individual add-in to extend review to that information. The intention was to obtain data directly from Revit documents and prepare them in a structure that could be queried through analysis tools, following a technical approach related to the extraction of data from federated models.
Separating data reading from processing
The first implementation focused on view information from linked models. Using the Revit API, the add-in identified related documents, queried their views, and organized the records for export to Parquet. I separated reading, processing, and writing so the project could later incorporate other review categories without rebuilding the entire workflow. Provenance and diagnostic identification mechanisms were also considered, which matters when a report brings together information from several models. Without them, a view with a repeated name could be difficult to associate with its source document.
What was tested
The initial functions ran in Revit 2022 with models from real projects. The next stages were intended to examine worksets, template use, family organization, and naming conventions. Incorporating the outputs into analysis processes already using Python, DuckDB, and Power BI for other audits was also considered. That expansion remained pending. Development stopped when priorities shifted to other initiatives; it remains a functional first stage rather than a tool that completed every planned check. The experience highlighted a useful distinction in BIM quality control. Reviewing what an element contains is not the same as reviewing how the file containing it was organized. The two perspectives can complement each other, but they require different data sources and evaluation rules.