Place Name Corrector
The 1870 Census for Washington County, TEXAS, has been corrected to Mississippi.
An example https://www.familysearch.org/ark:/61903/1:1:MXGF-2JW
@N Tychonievich please and thanks.
Best Answer
-
@Áine Ni Donnghaile That one is a little different from the auto-standardization issue we have been sending on to be fixed by the engineers. This looks to be an error created on the indexing side of things. Most likely, given the extent of the error, it happened when they were batching the census images in preparation for indexing. Short of opening a new indexing project, which practically never happens, we're probably stuck with this one. Fortunately, you have access to the images and can also edit the place to show properly when you use one of these as a source. Sorry!
If you're interested in knowing how to identify the ones that were caused by computer-driven auto standardization, take a look at the screen shot that you posted for the 1870 Census in this feed. The auto standardization ones always show both an (Original) Event Place and an Event place. Those are computer errors and the engineers are willing to tackle them.
0
Answers
-
@ Áine Ni Donnghaile Thanks. We'll add it to the auto-standardization report for the 1870 census.
1 -
Thank you.
0 -
Another one, please, @N Tychonievich
The 1840 census for White County, Tennessee, has been amended to be in Georgia.
https://www.familysearch.org/ark:/61903/1:1:XHYM-MVQ
There is no "original" reference on the extract, but the census page still shows Tennessee at the top.
Thanks.
0 -
Another one, please -
Middlesex, New Brunswick County, New Jersey is listed as Mercer County in the 1905 New Jersey State Census.
Example: https://www.familysearch.org/ark:/61903/1:1:KM4S-PFC
0 -
@N Tychonievich Thanks. I suspected that was the case on this one. Unfortunately, when I found it, the system would not allow me to submit an edit. I got a blank screen multiple times, on several different browsers. I'll try again another day.
0 -
As far as I can tell, this is an entire microfilm that has been auto-standardized to the wrong Szentbalázs:
https://www.familysearch.org/search/record/results?q.filmNumber=004620723
Individual example: https://www.familysearch.org/ark:/61903/1:1:X28M-B38
Baptism Place Szentbalázs, Somogy, Magyarország
Baptism Place (Original) Szentbalázs, Zala, Hungary
The correct place was disambiguated around the turn of the previous century as Zalaszentbalázs. The bot chose the Szentbalázs in Somogy, which was the one that got to keep the unmodified name. They're about 50 miles apart as the crow flies.
1 -
John Debour/Debois was born in the 1790s in Poland or Prussia or Germany, depending on the record, and based on border changes and wars. He settled in New York.
In this record, the 1908 death of his daughter, his place of birth is abbreviated to just Ger. 3 letters. The place name corrector has moved his place of birth to the Sudan.
https://www.familysearch.org/ark:/61903/1:1:2WFY-S5Q?id=GDL3-44P
0 -
Amanda Labour died on 30 December 1924, in Johnson City, New York. Her death is listed on the New York State Death Index. https://www.familysearch.org/ark:/61903/1:1:QG5Y-74K7
Place name corrector has put her death in Johnson, Kane, Utah.
0 -
Here is my great-great-grandmother Mary in the 1900 census: https://www.familysearch.org/ark:/61903/1:1:66VK-H731?id=KP7V-P18
If you look at the document itself, it clearly states that they lived in Rifle precinct. The house is currently in the city limits, but wasn't then. Still, it's only a mile or so from the post office in Rifle and is still lived in today.
The place name assigned is Parachute.
Here's one assigned to Carbondale, Colorado, when they live on Divide Creek (which is between Silt and New Castle). https://www.familysearch.org/ark:/61903/1:1:66VK-3FZV Even as the crow flies, the two places are 40 miles apart.
Here is the Coe family, who owned a ranch in Antlers, Colorado, which is halfway between Rifle and Silt. It's been assigned willy nilly to New Castle, even though the image says Antlers. I know where they lived. This is depressing. https://www.familysearch.org/ark:/61903/1:1:66VK-QJY8
Here is Olive Crook, The family lived in East Glenwood Springs. They have been given a place name of Cattle Creek: https://www.familysearch.org/ark:/61903/1:1:66VK-8PP8
David Heaton. Lived in Rifle. Assigned to Parachute. https://www.familysearch.org/ark:/61903/1:1:66VK-KRCB
Aunt Alta lived in Austin, Colorado, a place about 10 miles north of Rifle. https://www.familysearch.org/ark:/61903/1:1:66VK-D82G She has now been stuck in Carbondale, Colorado, a location 50 miles away.
0 -
Indexed baptisms from DGS 4594665, containing Roman Catholic records from Békéssámson, Békés county, Hungary, have been auto-standardized half to just the county of Békés, Hungary -- which is utterly useless, and not even correct until sometime after 1863, because it used to be in Csongrád county -- and half to the completely-incorrect "Sámson, Szilágy, Hungary", which is now Şamşud, Romania, 130 miles (as the crow flies) or a four-hour drive northeast of the correct place.
Records Search of the film: https://www.familysearch.org/search/record/results?count=20&q.filmNumber=4594665
Example of Event Place = just the county: https://www.familysearch.org/ark:/61903/1:1:XKBV-8BV
Example of Event Place = wrong country: https://www.familysearch.org/ark:/61903/1:1:XKBV-2HR
You'll also notice that half the baptisms have had a sex change (they're all now male), and the dates have been stripped of the month and day, rendering the index even less useful. I believe these changes are also the result of an automated updater run amok.
1 -
"St James, Essex, England" in Colchester has been mis-standardized to "Chichester St James, Chichester, Sussex, England, United Kingdom". Example index entry: https://www.familysearch.org/ark:/61903/1:1:M7HC-FDG
Thread about this error: https://community.familysearch.org/en/discussion/comment/426217#Comment_426217
(I originally posted this here on Feb. 6, but today "System admin" split it off on its own for some inexplicable reason: https://community.familysearch.org/en/discussion/comment/426286.)
1 -
I believe we're all caught up here now--reported the errors to engineering. Moving forward, please create a new Community post for each of these auto-standardization errors. Just makes my job a tiny bit easier. Thanks!
2 -
@N Tychonievich, I find it surprising that you prefer separate threads for these reports, but I'll follow your instructions. Is there a format or phrasing that you would find helpful? Also, people often report these as indexing errors or data errors or just "you all are crazy" errors. Would it be helpful for us to flag such reports in some way, and if so, how? Should we just at-tag you in a comment?
Relatedly: we've been concentrating on misstandardized places. Is there any hope of getting other recent bad changes to indexes (like the wholesale sex changes on Hungary Catholic Church baptisms) corrected? How or where can/should we report these?
1 -
Based on N Tychonievich's post here https://community.familysearch.org/en/discussion/116253/existing-historical-records-issues#latest, I propose to use "Inaccurate place standardization" as the title of future posts on this specific problem. Body will include the URL and a relevant snip.
2 -
@Julia Szent-Györgyi I prefer separate threads because it is easy for me to scan through the posts and see if I have responded to the ones that look like escalatable issues. Just my personal preference.
Usually folks use words in their titles when posting that make it pretty easy to see that they are reporting a potential issue/bug/mess with the system. We'll never get everyone to be uniform in how they post anyway. But, if you want to use key words like "Inaccurate place standardization", or "error message" or something along those lines, they might be picked up more quickly by the support agents who wander through Community looking for issues we need to get to engineering.
Wish I knew exactly what issues engineers accept and reject, but in general, let us know about things that seem to be a problem created by a computer rather than a problem created by less-than-diligent indexers. I have no problem with reporting something to engineering if I'm not sure. In every case, I need a specific example of the problem to send them. URLs are more useful to me that screen shots.
Hope that provides a little clarify.
2
