I did a survey of all 315 duplicate pairs and almost 270 of them had the
duplicate entity created with P428 Botanist Abbreviation set, which I
consider prima facie evidence that they came from the same source. For
some of the pairs, each entry was created within hours of each
<https://www.wikidata.org/w/index.php?title=Q21516236&action=history> other
<https://www.wikidata.org/w/index.php?title=Q21515511&action=history> by
Reinheitsgebot and Succu, one importing from Species Wiki and the other
importing from IPNI, with neither one checking for pre-existing entries.
Bots (and users) really need to take more care not to create duplicates
because they make future reconciliation tasks much harder for others.
Tom
p.s. Is there a way to get the edit history for an entity in JSON or some
other parseable format?
On Mon, Dec 28, 2015 at 2:29 PM, Tom Morris <tfmorris(a)gmail.com> wrote:
I think there are at least two uses for information
like this. Fixing the
actual errors is good, but perhaps more important is looking at why they
happened in the first place. Are there infrastructure/process issues which
need to be improved? Are there systemic problems with particular tool
chains, users, domains, etc? What patterns does the data show?
I've attached a munged version of the list in a format which I find a
little easier to work with and added Wikidata links
Looking at the 30 oldest entities, 22 (!) of the duplicates were added by
a single user (bot?) who was adding botanists, apparently based on data
from the International Plant Names Index
<http://www.ipni.org/ipni/authorsearchpage.do>, without first checking to
see if they already existed. The user's page indicates that they've had
2.5 *million* items deleted or merged (~12% of everything they've added).
I'd hope to see high volume users/bots/tools in the 99%+ range for quality,
not <90%.
One pair is not a duplicate, but rather a father
<https://www.wikidata.org/wiki/Q1228686> & son
<https://www.wikidata.org/wiki/Q716655> with the same name, apparently
flagged because they were both born in the 2nd century and died in the 3rd
century, making them a "match."
A few of remaining the duplicates were created by a variety of bots
importing Wikipedia entries with incompletely fused sitelinks (not terribly
surprising when the only structured information is a name and a sitelink).
The last few pairs of duplicates don't really have enough provenance to
figure out the source of the data. One was created just a couple of weeks
ago by a bot
<https://www.wikidata.org/w/index.php?title=Q18603442&action=history>
using "data from the Rijksmuseum" (no link or other provenance given),
apparently without checking for existing entries first. A few
<https://www.wikidata.org/w/index.php?title=Q16825734&action=history>
others
<https://www.wikidata.org/w/index.php?title=Q19619143&action=history> was
created
by Widar
<https://www.wikidata.org/w/index.php?title=Q19933200&action=history>,
but I can't tell what game, what data source, etc.
Looking at three pairs of entries which were created at nearly the same
time (min QNumberDelta), each pair was created by a single game/bot,
indicating inadequate internal duplicate checks on the input data.
It seems like post hoc analysis of merged entries to mine for patterns
would be a very useful tool to identify systemic issues. Is that something
that is done currently?
Tom
On Wed, Dec 23, 2015 at 5:05 PM, Proffitt,Merrilee <proffitm(a)oclc.org>
wrote:
Hello colleagues,
During the most recent VIAF harvest we encountered a number of duplicate
records in Wikidata. Forwarding on in case this is of interest (there is an
attached file – not sure if that will go through on this list or not).
Some discussion from OCLC colleagues is included below.
Merrilee Proffitt, Senior Program Officer
OCLC Research
*From:* Toves,Jenny
*Sent:* Tuesday, December 22, 2015 6:02 AM
*To:* Proffitt,Merrilee
*Subject:* FW: 201551 vs 201552
Good morning Merrilee,
You probably know that we harvest wikidata monthly for ingest into VIAF.
This month we found 315 pairs of records that appear to be duplicates. That
was a jump from previous months. I am not sure who would be interested in
this but Thom & I thought you might be. The attached report has 630 lines
showing what viaf saw as duplicates. So this pair of lines:
WKP|Q21518392 =998 $aCharles du Bois
Larbalestier$2WKP|Q21341290$3duplicate
WKP|Q21341290 =998 $aCharles du Bois
Larbalestier$2WKP|Q21518392$3duplicate
Shows that those two wikidata numbers are linked to one another by viaf.
I don’t think we expect you to do anything with this unless you find it
interesting. I suspect there are bots to clean this stuff up but maybe not.
--Jenny.
*From:* Hickey,Thom
*Sent:* Monday, December 21, 2015 9:47 PM
*To:* Toves,Jenny
*Subject:* RE: 201551 vs 201552
She probably would be interested.
--Th
*From: *Toves,Jenny <tovesj(a)oclc.org>
*Sent: *Monday, December 21, 2015 9:35 PM
*To: *Hickey,Thom <hickey(a)oclc.org>
*Subject: *RE: 201551 vs 201552
Exact same name + dates. Do you a list of them? Do you think Merrilee or
anyone would be interested?
*From:* Hickey,Thom
*Sent:* Monday, December 21, 2015 8:04 PM
*To:* Toves,Jenny
*Subject:* FW: 201551 vs 201552
Noticed WKP duplicates went way up
--Th
*From: *Jenny Toves <toves(a)orhddb01dxdu.dev.oclc.org>
*Sent: *Monday, December 21, 2015 5:12 PM
*To: *Hickey,Thom <hickey(a)oclc.org>rg>; Toves,Jenny <tovesj(a)oclc.org>
*Subject: *201551 vs 201552
REPORT for records
Changed 13.51%: geographic 3369217.0 -> 3824513.0
Change in % of 8: NLR at_least_one_match 16% -> 24%
Changed 19.83%: NLR all_matches 181437.0 -> 217423.0
Change in % of 88: NLR with_bibs 0% -> 88%
Changed 17.99%: WKP geographic 2529990.0 -> 2985194.0
Changed -19.95%: WKP corporate 369224.0 -> 295579.0
REPORT for matches
Changed 12.70%: exact corporate name 1021239.0 -> 1150899.0
Changed 14.29%: XR viafid 7.0 -> 8.0
Changed -10.42%: XR expression title to sibling 48.0 -> 43.0
Changed -16.16%: PTBNP forced 229.0 -> 192.0
Changed -37.50%: NSZL forced 8.0 -> 5.0
Changed 38.46%: NLP suggested 13.0 -> 18.0
No longer zero: NLR standard number 0.0 -> 21479.0
No longer zero: NLR exact title 0.0 -> 5166.0
No longer zero: NLR partial date and partial title 0.0 -> 618.0
No longer zero: NLR name as subject 0.0 -> 62.0
No longer zero: NLR partial title and publisher 0.0 -> 88.0
No longer zero: NLR title 0.0 -> 5093.0
Changed -47.66%: NLR forced single date 37125.0 -> 19430.0
Changed 14.29%: NLR viafid 14.0 -> 16.0
No longer zero: NLR partial date and publisher 0.0 -> 15894.0
No longer zero: NLR joint author 0.0 -> 5228.0
Changed -14.49%: LC suggested 7594.0 -> 6494.0
Changed 33.33%: CYT viafid 12.0 -> 16.0
Changed -21.08%: NLA forced 223.0 -> 176.0
Changed 233.33%: LNL forced 3.0 -> 10.0
Changed 12.50%: NLB viafid 8.0 -> 9.0
Changed 16.67%: NLB ngram corporate name 6.0 -> 7.0
Changed 25.71%: VLACC forced 35.0 -> 44.0
Changed 19.13%: DNB exact corporate name 315872.0 -> 376304.0
Changed 14.29%: DNB expression title to sibling 7.0 -> 8.0
Changed 16.67%: BNF expression title to sibling 6.0 -> 7.0
Changed 15.91%: ICCU forced 44.0 -> 51.0
Changed 25.54%: NTA forced 9699.0 -> 12176.0
Changed 28.62%: WKP exact corporate name 224787.0 -> 289112.0
Changed 23.73%: WKP longer corporate name 76057.0 -> 94106.0
Changed 584.78%: WKP duplicate record 92.0 -> 630.0
Changed -18.92%: EGAXA forced 37.0 -> 30.0
REPORT for tags
Changed 11.56%: NSZL work links (993) 225.0 -> 251.0
No longer zero: NLR wrote about (955) 0.0 -> 106.0
No longer zero: NLR bibs (999) 0.0 -> 108202.0
No longer zero: NLR was a subject (960) 0.0 -> 16423.0
No longer zero: NLR relator code (941) 0.0 -> 103950.0
No longer zero: NLR language of work (940) 0.0 -> 108193.0
No longer zero: NLR issn (902) 0.0 -> 34.0
No longer zero: NLR bib title (910) 0.0 -> 107895.0
No longer zero: NLR joint corporate author (951) 0.0 -> 24235.0
Changed 146.67%: NLR compared (996) 27448.0 -> 67705.0
No longer zero: NLR rectype + biblvl (944) 0.0 -> 108194.0
No longer zero: NLR country of publication (922) 0.0 -> 108169.0
No longer zero: NLR publisher (921) 0.0 -> 93904.0
No longer zero: NLR isbn (901) 0.0 -> 78978.0
No longer zero: NLR publisher id (920) 0.0 -> 78978.0
Changed 50.05%: NLR matched (998) 19864.0 -> 29806.0
No longer zero: NLR name from statement of responsibility (930) 0.0 ->
72478.0
No longer zero: NLR noise title (912) 0.0 -> 3543.0
No longer zero: NLR lc class number (942) 0.0 -> 1.0
No longer zero: NLR joint author (950) 0.0 -> 69048.0
No longer zero: NLR was a subject (969) 0.0 -> 115.0
Changed -14.29%: XA work links (993) 7.0 -> 6.0
No longer zero: SRP work links (993) 0.0 -> 1.0
Changed 22.50%: BNL work links (993) 551.0 -> 675.0
Changed 11.16%: WKP auth title (919) 45779.0 -> 50890.0
Changed 12.56%: WKP noise title (912) 8249.0 -> 9285.0
_______________________________________________
Wikidata mailing list
Wikidata(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata