FTA exporting Lost Cousins data

However Peter you are missing the point that users don't have to add those tags to use the program or the website, it's a little extra "nice to have". If they do use it it makes the reports more intelligent and shorter next time. However a user such as yourself with NO census tags and no Lost Cousins tags could use the program today and auto search FamilySearch get the 1881 census record for someone in their tree and add it to the Lost Cousins website in a few seconds.

In fact by adding a couple of FTAnalyzer screenshots to your document you could make it even quicker for the user as they wouldn't even have to do the search on FamilySearch one double click would do the search for them. So they would cut down the time needed to enter a new family into Lost Cousins even more. Ie: instead of showing the user how to slowly manually search FamilySearch they could auto search FamilySearch using FTAnalyzer.

This is something you could try with your own tree right now it takes seconds to do and shows the power of being able to use a program to quickly add Lost Cousins data. In fact I'll go and change the default search to FamilySearch so that your users have one less step to do.

About the only thing that could be done to improve this process would be to have FTAnalyzer auto fill out the Lost Cousins website for them.

 

So, as I said

I have just downloaded the data from LC and the 1881 report from FTA into excel.

After removing the leading space from the LC data census ref (use the function TRIM), and filtering on the same census ref in both files, I now have a perfect match everytime. I don't need to introduce a new field or worry about name changes or dates.

So here is my data from LC


And from FTA


So its a good example of where the names are completely different due to misspellings.

Alexander, Peter mentioned this

Have you considered loading a 2nd file to supplement a gedcom? People who record LC facts don't have to do this, but for those that haven't yet?
Instead of listing people the new report could say, These census ref's were not found on LC, and then list the relatives for people to load at LC?

 

I've used FTA myself - that's how I know how much work I'd have to do before I could begin using it to pinpoint and relatives I've inadvertently omitted. I don't have CENS facts, I don't have LostCousins facts, I haven't even recorded census references in my family tree - perhaps I should have done, but it isn't the way I started using Genopro in 2002, and it's hard to imagine when I'm going to find the time to go back and make thousands of changes. It's not that I don't record the information at all - it's just that it isn't in my family tree.

Because there are a lot of people like me - including many who only have an online tree - I've been trying to explore whether there is some way in which members can identify the relatives they need to enter without having to first enter additional information for every individual. If the aim is to encourage more people to enter more relatives on their My Ancestors page then the easier we can make it for them the better.

If what I've suggested isn't possible, that's not a problem - but you can't blame me for asking. I spend most of my life trying to encourage members to enter more data, and I'm always trying to find ways to make it that little bit easier - for example, on Friday I posted three new Getting Started guides on the Help & Advice page to make it easier for members who haven't entered any relatives at all to get going.

Postscript: the system I specified back in 1977 (for the retail chain I was then working for) used a very similar matching procedure to the one I outlined earlier in order to reconcile supplier accounts - and we were faced with very similar problems. It would have been so easy if there had been a unique reference we could match on (an invoice number, for example), but in many cases there wasn't, and suppliers weren't going to change their systems just to oblige us.

 

A day or two ago I asked Red Dragon about this another thread. Tim replied in Red's absence, but even after following his advice I was still unable to generate a list of relatives in my tree who were (or may have been) alive in 1881. Maybe I didn't click the right buttons, but clearly I can't explain to other members how to use FTA in this way if I'm unable to do it myself.

Could it be because there's nothing in my Gedcom to tell FTA which country my relatives lived in (I only ever mention the country when it isn't England)? Is there a setting in FTA that would overcome this?

 

Yes I considered it and said it would work best if there was the option for users to enter a reference ID against a name as then you have a specific thing to match against. At present most people don't record a census ref in their file (you only recently started doing that as have others), my own file has surprisingly few census refs and most need manual tweaking to be consistent enough to match. Thus for the vast majority of my own records I'd struggle to match anything. Thus it would be a right royal pain to implement.

There is a danger of several things getting conflated here. There are very different and separate issues being discussed intertwined...

1) Who is alive on the 1881 census can we report that?
2) Viewing who has already been entered on the 1881 census
3) Comparing who has been entered on the Lost Cousins website
4) Importing data from Lost Cousins website to update the users tree
5) Automatically adding people from your tree to the Lost Cousins website

1) Yes this is possible this was the very first report added to FTAnalyzer over 8 years ago back in the days when it was a java webservice.

2) Yes this is possible it requires the users tree program to support the GEDCOM standard CENS tag - not all programs or perhaps more accurately not all users do this.

3) This requires lists to be sorted in the same manner on the website and FTAnalyzer this has largely been achieved but could be tweaked to be better. Note this typically relies on having common data in both reports eg: user recording census references in their tree. Or having an individual reference ID on Lost Cousins website as I suggested, either way some tweaks are required to make the sorting similar enough to be usable.

4) This requires everything in 3 and then a whole lot more, it requires an idealised tree where the names are tidied up the dates are fixed and the user has done a lot of manual work, it then requires the users program to actually support importing and merging data and finally and most important of all it requires the user to implicitly trust that importing data won't damage their tree. This one is a dead duck as I fundamentally don't like the idea of tampering with a users tree there is far far too much room for things to go wrong and blame to be attributed.

5) This is different it doesn't require anything more than we have at present, before entering data onto Lost Cousins the user needs certain minimum information. FTAnalyzer could easily create a report of data that the user had yet to enter by REQUIRING census references that completely changes the ball game of 3 above in that it gives a reference to match on. It would then be fairly simple to generate a file that the user could upload to Lost Cousins website and would contain new entries not on the website already. By having requirements before export only entirely valid records would be exported. This would enable users to potentially upload hundreds of new relatives in a few seconds. This would probably require new code on the Lost Cousins site.

So very different subjects that were getting mixed up. The program can already auto search FamilySearch and allow users to see who they can quickly add to Lost Cousins. Adding number 5 as an option and allow users to batch upload people for whom they have a census ref, age, birth location etc. would eliminate at a stroke the major objection people have to using Lost Cousins ie: they've already entered the data why do they have to enter it again.

 

It could be missing countries, it depends on how you record locations. I did just produce an example file of your John & Emily Calver family for the 1881 census for you to see how it works. That should be available to you in the PM section.

FTAnalyzer can interpret countries, and I can add more code to interpret further if you have a particular way of recording things. Would it be possible to export a GEDCOM of say just John & Emily as an example so I can see how your location data is being treated. Perhaps I can improve the country handling.

 

Many people use the optional section of the form to correct the census information - when they do, it's the corrected version that shows up on their My Ancestors page.

 

That's even better then, the census ref's will align and the corrected info will also match.

 

Sorry Peter, this is how you generate a list of people who were or could have been alive in 1881.

After loading your gedcom click on the Census Tab.

The Census Date defaults to UK 1881 so now just press the "Show Missing from Census" button.

All the people listed here were or could have been alive for the 1881 census and you don't have a recognised census reference for them.

 

Thanks, Tim - I had tried that but when I did only 3 individuals were displayed (out of thousands).

However, I've just realised why - it's because the wrong root person is selected. Can you remind me how to change the root person in FTA?

 

Yes, on the Individuals Tab, right click on the person you want as Root, and then left click to set it.

 

Good to see you've got it working.

 

Thank you both.

Now that I can see the results it has highlighted another anomaly in my tree - I don't routinely enter approximate dates of events, which perhaps I ought to - but being optimistic I assume that I'm going to discover the register entry fairly soon. As a result I'm getting people in the list for whom no birth or death dates are included, even though there is other data in the tree that would exclude the possibility of them being on the 1881 Census. For example, one was the daughter of someone born in 1892; another married in 1722.

Rather than make the program more complex, could we perhaps have an option to exclude people for whom no birth or death date is known? When I find someone on the census I do enter a birth year, even though it might be a few years out, so in my tree the absence of a birth year usually indicates someone who hasn't been found on any published census (usually because they were born too early or too late).

 

The "Loose" Births report on the births deaths tab would show you what information you already have in your tree that could narrow down the birth dates of people with unknown births. However from the sound of it this might be a LOT of people.

I already filter out people who are too old on a census date (the default is 90+) so the filtering on unknown death is already dealt with. I could add an option to exclude those with unknown birth dates as a work around for the issue you are currently facing although you are probably right you are probably better entering approximate births eg: for daughter of person born in 1892 you could assume mother had to be 16+ (or perhaps 14+ if you prefer) in which case the daughter would have a birth of AFT 1908. The one married in 1722 again could assume was 16+ at marriage so birth becomes BEF 1706.

It's actually a really good tool when working with trees to enter such narrowing down data as it really helps clear up timescales eg: you know the first one is well in statutory and even likely to have a birth registration including mothers maiden name (from 1911 in England as I understand it), whereas the second one you know will only appear in OPRs. Having loads of unknowns means you can struggle to see the woods for the trees.

Unlikely this would be the case for yourself but I've known of so called brick walls to disappear when people have adopted approx dates as suddenly they realise for instance that a particular family must have been alive on a census they put more effort into it and the "brick wall" shatters.

Using lots of unknown dates makes your research harder. As you say the data is there in your tree, for some people just clearing up this unknowns confusion suddenly makes things fall into place, eg: seeing that someone couldn't have been the daughter of X and they must have been the grand daughter. Etc.

 

Version 3.6.1.3 now available that implements filtering to exclude people with unknown births.