metadata at icpsr sanda ionescu, icpsr. metadata at icpsr -catalog records- created by data...
TRANSCRIPT
Metadata at ICPSR
Sanda Ionescu, ICPSR
Metadata at ICPSR-Catalog Records-
• Created by data processors, who fill out a Web-based form:– Fixed fields,– DDI 2.x and 3.0 compatible
Metadata at ICPSR-Catalog Records-
Metadata at ICPSR-Catalog Records-
• Review and approval by metadata specialist• Stored in ORACLE database• Exported from database to DDI 2.1 XML• XML files stored on server (file system)• HTML and PDF presentation created dynamically
(through XSLT stylesheets) at user request• HTML presentation for viewing only; PDF is
downloadablehttp://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/08589.xml
Metadata at ICPSR-Catalog Records-
• DDI-XML files searched by field from home page to retrieve studies (Inktomisearch)
Metadata at ICPSR-Codebooks-
• HERMES – in-house automated process to generate (most of) the study distribution package:– Input:
• SPSS system or portable• Optional pre-formatted (question) text file
– Output: • Full suite of statistical formats (setups and system)• ASCII data file• DDI 2.1 file with frequencies and question text if available
Metadata at ICPSR-Codebooks-
• DDI 2.1 file may be converted to PDF to generate – An “ICPSR” codebookhttp://www.icpsr.umich.edu/cgi-bin/bob/archive2?study=4699&path=ICPSR&docsonly=yes
– Part of the publicly distributed codebook as other non-DDI resources may be incorporated
http://www.icpsr.umich.edu/cgi-bin/bob/archive2?study=4512&path=ICPSR&docsonly=yes
• In some instances a DDI-based codebook will not be generated
http://www.icpsr.umich.edu/cgi-bin/bob/archive2?study=9522&path=ICPSR&docsonly=yes
Metadata at ICPSR-Codebooks-
• The DDI 2.1 file with variables description– Is archived– Is downloaded into the Social Science Variables
Database (SSVD)
Metadata at ICPSRSocial Science Variables Database
• Also built in ORACLE, but currently a separate entity, with links to studies’ and series’ descriptions.
• Includes variable-level metadata.• Is DDI 2.x and 3.0 compliant (input and output)• Will enable variable-level searches across studies
and series of studies (simple SQL queries - retrieve matches, do not infer relevance)
Integrating DDI 3 into ArchivesSRO-ICPSR collaboration project
Common RELATIONAL DATABASE model for data documentation- Compliant with DDI 3.0 -
Common RELATIONAL DATABASE model for data documentation- Compliant with DDI 3.0 -
Blaiseoutput
Blaiseoutput
SAS/SPSS/Stata files
SAS/SPSS/Stata files DDI 2.xDDI 2.x DDI 3.0DDI 3.0Other…Other…
Client Applications… Web Applications…
SRO: ICPSR:
ICPSR: Variable-level Search
ICPSR projects will be able to use documentation generated by SRO projects…
Metadata at ICPSR-Online analysis-
• Survey Documentation and Analysis (SDA) – Approx. 475 studies – data and documentation in
proprietary format (ddl), DDI 2.x-compatible.– Nesstar - used only as a “test” (currently not in
production mode)https://www.icpsr.umich.edu/ICPSR/access/sda.html
Metadata at ICPSR
• Other study documentation– Questionnaires– User guides– Data definitions
Distributed in machine-readable, but non-searchable formats – PDF, ASCII, Excel, etc.
http://www.icpsr.umich.edu/cgi-bin/bob/archive2?study=4512&path=ICPSR&docsonly=yes