Earlier this month, DDI members voted on the Structured Data Transformation Language (SDTL). Nineteen of the 28 eligible voting designated member representatives voted, with all responding “YES, the validity and usefulness of Version 1.0 of SDTL has been demonstrated and it should now be accepted as a part of the DDI standard.” The vote passed with a two-thirds majority.
SDTL is an independent language for representing data transformation commands in statistical analysis packages, such as SPSS, Stata, SAS, R, and Python. Commands like RECODE, MERGE FILES, and VARIABLE LABELS are rendered in a structured format (JSON, XML, RDF) that is easy for machines to read and process. Command scripts translated into SDTL produce variable-level data transformation histories, which can be translated into natural language. SDTL can be added to Data Documentation Initiative (DDI) and other metadata standards for use in data catalogs, codebooks, and other documentation. SDTL was developed by the Continuous Capture of Metadata for Statistical Data (C2Metadata) Project, which was funded by National Science Foundation grant ACI-1640575.
As a next step, the DDI Technical Committee will incorporate corrections noted in the Public Review and prepare the specification for publication.
Thanks to everyone who contributed to the development of SDTL, especially George Alter (Principal Investigator of the C2Metadata Project), the C2Metadata Project members, the DDI SDTL working group, and the DDI Technical Committee.
More details about SDTL:
- SDTL User Guide: http://c2metadata.gitlab.io/sdtl-docs/master/
- Introduction to SDTL Webinar:https://youtu.be/62dLnRkwQUg
- Introduction to SDTL Paper: edu/handle/2027.42/156015
- Overview of the C2Metadata Project: href="https://deepblue.lib.umich.edu/handle/2027.42/156014
- SDTL Working Group: https://ddi-alliance.atlassian.net/wiki/spaces/DDI4/pages/899547182/SDTL+-+Structured+Data+Transformation+Language