Friday, January 23, 2015

Darned Records: Mickey Mouse’s Vitals

If you thought Mickey Mouse was born at the Disney Studios in California in 1928, the Disneyland Hospital in Chicago, Illinois may beg to differ. He was born there in 1918. And who knew that he’s dating a mouse with the same name as his mother. The registrar, perhaps fearing his job, chose to remain anonymous. (Thank you, Sarah Stoddard, for providing this birth certificate.)

A parody of a birth certificate of Mickey Mouse

Next is Mickey’s obituary. Mickey Mouse is very much alive, so Ancestry.com’s record of his obituary is very much in error.

Ancestry.com has erroneously posted an obituary of Mickey Mouse.

You might ask yourself how this happened.

Over the years, companies have tried to develop automated algorithms that allowed machines to read and understand text. This is more than OCR. It is more than Word putting blue wavy lines under common grammatical errors. It is called natural language processing. And it is really, really difficult. Mickey’s obituary is a case in point. The software took an obituary of Ralph Kent and turned it into an obituary of Mickey Mouse. I would have provided a hotlink to the botched obituary, but, as I’ve pointed out before, many of the links in the Ancestry.com United States Obituary Collection are broken.

Lessons learned:

  • You can’t safely use a record unless you understand it. In this case, knowing that a machine used natural language processing to extract information from free form text is key in understanding its evidentiary value.
  • Record abstracts are suspect. Like any derivative, consult the original.

Yes, “Records Say the Darnedest Things!”


Sources

     Birth certificate image: Cook County, Illinois, birth certificate no. 6153 (1 January 1918), Mickey Mouse, County Clerk, Chicago; image provided by Sarah Stoddard. This birth certificate is a parody of Mickey Mouse. See “Fair Use,” Wikipedia: The Free Encyclopedia (http://en.wikipedia.org/w/index.php?title=Fair_use&oldid=641591802#Fair_use_and_parody : 8 January 2015, 16:06 revision), “Fair Use and Parody.”
     Obituary image: “United States Obituary Collection,” automated extract, Ancestry (http://search.ancestry.com/cgi-bin/sse.dll?indiv=try&db=web-obituary&h=15878266 : accessed 10 January 2015), search for Mickey Mouse; citing Boston Herald, online edition (http://www.bostonherald.com/news/obituaries/general/view.bg?articleid=1031680&format=text : 16 September 2007); the link did not work on 10 January 2015.

3 comments:

  1. Ancestry is becoming worse and worse about "facts". Facts gleaned from unsourced family trees are not facts. Family Data "facts" are a joke.

    ReplyDelete
  2. Years ago my mother paid a fair sum of money to some professional genealogists. She eventually fired them when she realized they were having a difficult time reporting back accurately information that she had given them in the first place. Ancestry cites those same records as The Millennium Research File. I have to admit I laugh whenever I see that in a citation.

    ReplyDelete

Note: Only a member of this blog may post a comment.