User:Jheald/BHL

cf Wikidata:WikiProject BHL

Undertaken / underway

Authors (216374)

• • Originally added using author-string
    • Need to fix author-strings contaminated with eg "Edition: 1st ed. 1st thousand Hansen" : 3897 contaminated tinyurl.com/ychx3sfo
  • Some with matched BHL ids replaced with author - 7932 tinyurl.com/yd4o6dx7
    • Check leftovers, to see which BHL ids matched to items could not be matched to author-strings
    • Systematically try to match to VIAFs; eg via OCLC ids; fall-back: BHL ids with dates
      • Comprehensive list of BHL IDs would be useful

Type (129278)

• • 2120 identified as periodicals, based on KW: tinyurl.com/ydbpgppg
    • See d:Wikidata:WikiProject Periodicals
    • Classify into types of periodicals / ? serials
      • Current classification: [1];
        • cf LCGFT Property_talk:P4953/entries#Informational_works -> Serial publications (rather limited);
        • Also User:Jheald/aat/full -> serials (publications)
    • Need start/end date
    • Break out volumes
  • 178 with more than 10 IA scans tinyurl.com/yazg4ya8 -- (mostly) these are also periodicals
    • But some might be book series
    • Likely to be more to come, because IA links weren't added to items with 4-digit (ie year?) volume identifiers, eg Q51382159
  • Other types that shouldn't be editions
    • eg: Tech reports, short papers <--  ? get page count from IA ? And/or identify series ?
    • Some of these may be revealed through links to series

More fields

• Edition (5354)
  • Added to 4534 tinyurl.com/yctrdlm7 (easy cases)
    • 820 remaining
• Publication place (117826)
  • 6046 distinct places identified
    • 1651 matched with OpenRefine --> 43,590 statements to be added (102,117 volumes)
      • • query now finds 43,575 tinyurl.com/yawfkasg
      • Should possibly be further checked, -- some errors, eg: Augusta (Maine) -> Augsburg (in Q51474060) x36, Albany (New York) -> Auckland (Q51457157)
        • Where there is a qualifier (eg state abbreviation, etc), check these match
        • Where there is no qualifier, look particularly closely (eg Augusta, Cambridge)
        • Look at geography of languages /tinyurl.com/y8ta3bmw
      • Add sine loco (Q11254169) where appropriate
• Publisher (115581)
  • Added using <somevalue>, stated as (56,556 distinct)
    • Need to try to match some of them to items

IA links

• • Links need to be added where there is more than one scan-item per BHL item.
    • Underway: links being added where there are easy volume numbers
    • Need to see what is left over --> other items with volume numbers not so easily interpreted
      • Do we now have some of the volumes already ?
  • Items with matching volumes from multiple sources 103 tinyurl.com/y8c4y2bt
    • Some need to have volume idenfiers added; ie multivolumes or volumes with parts
    • others are genuinely multiple copies of the same volume from the same institution

To do

From IA pages

eg [2] :

• Get number of pages
• Get language
• Double-check volume #
• Get OpenLibrary edition / OpenLibrary work

From BHL

• Get BHL IDs, standardised names, list of Title IDs / Vol IDs.

• Check individual book pages for genre --> eg 'journal' [3] (not always in keywords)
  • Move volume information for journals to separate items (but first check v. carefully for journal duplication)

• Other fields on BHL pages ?
  • language
  • series
  • related works
  • ... ?
  • role information for contributors

From OCLC pages

• OCLC ids for BHL books in WD with unmatched authors: tinyurl.com/y7ygbxma (42,143)
• Get author VIAFs (& compare with BHL IDs) -- underway
• Get country of publication

Constraint violations

• Multiple copies with same OCLC [4] (1395)
• Also check multiple entries with same IA link -- eg monographs appearing as complete issues of serials

On page

• Fix descriptions

BHL fields / progress

from BHL_data.txt:

• KW -- keywords (444907)

-- could identify most common & match -- suitable for 'main subject' ? Or use another property for faceted description?

• AU -- author (216374)

-- authors with BHL / Qid match already marked, except where multicle candidates not distinguished -- extract more BHL ids --> more books --> VIAFs --> Qids -- add BHL id as qualifier ?

• ER -- end record (129278)

• UR -- URL (129278)

        --> BHL identifier --> already added

• TY -- type, ie "Book" or "Periodical" (129278)

-- periodicals now marked --> check titles for more detailed, eg Journals -- non-periodicals --> should mostly be marked as edition or volume

• TI -- title (129278)

       -- already added

• PY -- Publication year (128661)

       --> already added

• CY -- City (of publication) (117826)

   --> identify Q-numbers

• PB -- Publisher (115581)

       --> in progress

--> identify Q-numbers (with city?)

• N1 -- notes (92954)

       -- should be added

• VL -- volume (91243)

• SN -- shelf number (31813)

     --> ignore ?

• ET -- edition (5354)