picture thisfiles.archivists.org/conference/dc2014/materials/p16-gignacs.pdf · drupal content •...

1
PICTURE THIS : Migrating Metadata in Visual Materials Collections 1. Evaluate existing metadata structure Reviewed our current metadata, checking it against the Dublin core standard Added “Location”, “Format”, “Collection”, and separated “Keywords” from existing “Photo Description” field and “image title” 2. Select a new content management system Drupal is a very flexible and powerful platform which was necessary for AIP’s diverse needs and audiences Our internal web development team could handle the configuration and customization that Drupal required 3. Evaluate current metadata quality; pricing structure Cleaned up fields before migration Made sure only LOC-authorized terms were included in our “Keywords” and “Location” lists Verified and deleted all duplicate name authority, “keyword”, “location” entries New fields: “Location”, “Format”, “Collection”, “Keywords”, “Photo Description”, “Image Title” Separated “Photo description” from “Keywords” field 4. Design a CMS that fits your needs Redesigned our cataloging and digitization workflows on the new CMS New image uploading process designed as a “one-stop shop” 5. Improve user search capabilities Chose Drupal modules that would enable advanced search feature Grouped images into themed galleries “Keywords”, “Location”, “Date”, and “Categories” fields were hyperlinked under each image record, bringing users to an aggregated page of results 6. Formulate a migration strategy Updated new metadata fields (20,000/25,000) before migration Created an Excel spreadsheet with all the data fields and how they would map from Access database to Drupal content Worked out what needed to be translated (comma- delimited lists of keywords, personal names, etc.); wrote functions to translate into new format for content management system 7. Migrate data Migrated a small test set of records (A through B); verified the migration results (approx. 1,000 records) Evaluated results by testing: Batch upload > testing > evaluate, then repeated process Eventually, completed final migration of 25,000 image records Savannah Gignac | Photo Librarian | Emilio Segrè Visual Archives | American Institute of Physics | One Physics Ellipse | College Park, MD 20740 Top Tips from Lessons Learned Final migration means final: Be clear about when the final migration is happening. If any metadata is updated in the interim it will be lost by next migration. Never update in between migrations. Document tasks and team progress: We had a Drupal “Basecamp” in which we communicated goals, progress, and deadlines. Use Google, a personalized wiki, or blog to record the choices made. If not, you will forget the thousands of steps you took during the migration process. Write down or record all verbal decisions: Phone and Skype conferences are good tools—but miscommunications can happen in big projects when deadlines and to-dos are based on memories, not print. Test user accounts from every angle: Be sure to try as many different user scenarios for searching and ordering as possible. There are exceptions to the rule you need to plan for— before they occur. Our Timeline 2013 2014 JAN Research begins JULY Development begins MAR Planning begins OCT First migration DEC Final migration MAY Final stage: Launch https://photos.aip.org

Upload: others

Post on 13-Oct-2020

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: PICTURE THISfiles.archivists.org/conference/dc2014/materials/P16-GignacS.pdf · Drupal content • Worked out what needed to be translated (comma-delimited lists of keywords, personal

PICTURE THIS: Migrating Metadata in

Visual Materials Collections

1. Evaluate existing metadata structure

• Reviewed our currentmetadata, checking it against the Dublin core standard

• Added “Location”, “Format”, “Collection”, and separated “Keywords” from existing “Photo Description”fieldand“imagetitle”

2. Select a new content management system

• Drupal is a very flexible and powerful platform which was necessary for AIP’s diverse needs and audiences

• Our internal web development team could handle the configuration and customization that Drupal required

3. Evaluate current metadata quality; pricing structure

• Cleanedupfieldsbeforemigration

• Made sure onlyLOC-authorizedterms were included in our “Keywords” and “Location” lists

• Verified and deleted all duplicate name authority, “keyword”, “location” entries

• Newfields: “Location”, “Format”, “Collection”, “Keywords”, “Photo Description”, “Image Title”

• Separated“Photodescription”from“Keywords”field

4. Design a CMS that fits your needs

• Redesigned our cataloging and digitization workflows on the new CMS

• New image uploadingprocess designed as a “one-stop shop”

5. Improve user search capabilities

• Chose Drupal modules that would enableadvancedsearchfeature

• Grouped images into themedgalleries

• “Keywords”, “Location”, “Date”, and “Categories” fields were hyperlinked under each image record, bringing users to an aggregated page of results

6. Formulate a migration strategy

• Updated new metadata fields (20,000/25,000) beforemigration

• Created an Excelspreadsheetwithallthedatafieldsand how they would mapfrom Access database toDrupal content

• Worked out what needed to be translated (comma-delimited lists of keywords, personal names, etc.); wrotefunctions to translate into newformat for content management system

7. Migrate data• Migrated a small testset of records (A through B);

verifiedthemigrationresults(approx.1,000records)

• Evaluatedresults by testing: Batch upload > testing > evaluate, then repeatedprocess

• Eventually, completedfinalmigration of 25,000 image records

Savannah Gignac | Photo Librarian | Emilio Segrè Visual Archives | American Institute of Physics | One Physics Ellipse | College Park, MD 20740

Top Tips from Lessons Learned

• Finalmigrationmeansfinal:Beclearaboutwhenthefinalmigrationishappening.Ifany metadata is updated in the interim it will be lost by next migration. Never update in between migrations.

• Documenttasksandteamprogress: We had a Drupal “Basecamp” in which we communicated goals, progress, and deadlines. Use Google, a personalized wiki, or blog to record the choices made. If not, you will forget the thousands of steps you took during the migration process.

• Writedownorrecordallverbaldecisions: Phone and Skype conferences are good tools—but miscommunications can happen in big projects when deadlines and to-dos are based on memories, not print.

• Testuseraccountsfromeveryangle: Be sure to try as many different user scenarios for searching and ordering as possible. There are exceptions to the rule you need to plan for—before they occur.

➔ ➔

Our Timeline2013 2014

JANResearch begins

JULYDevelopment

begins

MARPlanning begins

OCTFirst

migration

DECFinal

migration

MAYFinal stage:

Launch

https://photos.aip.org