Mightier Acorns

The Callan Name Study keeps growing, and as it does, I intend to tell you about it. If you know of anyone with a Callan surname (no matter what spelling they used), take a look at the master list (a spreadsheet you can find here or on WikiTree) and see if I have them in the database. If not, you can add them with this form!

I Have Lists of Names…What’s Next?

Over the past couple of months, I’ve experimented with Google’s Gemini with variations of this prompt:

Generate a .CSV file listing every individual in the 1850 US Census whose 
surname is a variation of “Callan” (using Soundex to generate the 
spelling variations) and provide columns for the following information: 
Surname, First name(s), age, year of birth (calculate this by subtracting the age value from the date of enumeration for the census page), sex, place of birth, head of the household, and columns for the State, County, and town or city of the census place for that person. 

After tinkering with the wording and figuring out what Gemini could and couldn’t do, I ended up with several spreadsheets – one for each available census year in each of the English-speaking countries (U.S., Ireland, Australia, and Canada, plus individual census records for England, Wales, Scotland, and Northern Ireland). I was able to combine them into two tabs on this Proton spreadsheet. (The first tab is a backup of my One Name Study, which is on Google Docs.)

What I don’t know is “how complete is this?” The results seem complete, but I don’t have an easy way to perform any quality control. At some point, probably during my summer break, I intend to run this experiment again and try a few tricks to determine if the results are different. If I take the outputs from the first and second sets of results and use the “remove duplicates” utility, I should be able to see whether they match perfectly or not. And if not, I may need to run the experiment several times until I can determine I’ve gotten “everybody” or as close to that as I’m likely to get.

I’ve also tried a few different methods to get Gemini to compare the results across multiple decades to see if it can identify the same individual in different census years, but for now, that task seems better suited to a human being.

Apples and Oranges

I also ran the prompt to get CSV files for the 1790-1840 U.S. Census records. For those who may not be aware, 1850 was the first U.S. Census to record the names of each person in a household. Before that, only the Head of Household was named, and everyone else was counted by gender and age ranges (which differed from decade to decade).

I haven’t included those results in the US tab of the spreadsheet because their format is so different, but I need to figure out a way to compare all of these results. But, since there is even less identifying information to work with (ie, we don’t know the age of the head of the household in these records), it’s going to come down to human judgment making the call.

And I suspect we will have to use a very manual, very labor-intensive process to get all of these individuals and families mapped into Ancestry/WikiTree/FamilySearch/etc.

What’s In A Name II: Electric Soundex Boogaloo

I don’t know yet how much of a problem this might be, but here’s what I get when I ask for the Soundex variations of Callan:

Phonetic Soundex Key: C450 (Groups primary late-18th-century spelling variations including Callan, Callen, Callin, Callon, Calan, Calen, and Callam).

I know that there are several other variations (like “Callaghan” and the occasional “O’Callan” or “McCallen”) that this leaves out. I may need to go look again and see if there is a robust One Name Study for variations of Callahan, or of the similar-but-distinct “Collins,” and ask if they’ve seen any consistent data that shows our surnames either come from the same source, or for different families.

Summer Break

By the time you read this post, I should be back from traveling, but before I get back to regular posting, I’m going to give myself a bit of a summer break. I expect I’ll still obsessively pore through all of this data and write stories about what I find interesting, but look for those stories to start appearing again after August starts to wind down.