Are you Using WikiTree Tech to assess profiles for Genealogically Defined completeness? ?

Question

In 2017 I introduced Robert Charles Anderson's concept of Genealogically Defined Profiles to WikiTree (See Here).

My goal at the time was to offer a way for a WikiTreer to assess a profile for a basic level of completeness while providing real evidence to support their statements.  

This was a kind of beginning or basic standard that encouraged sourcing all relationships and vital information.  The idea was to promote evidence based statements for all relationships established.

New WikiTree Technology has grown quite a lot since that time that we can almost automate that assessment.   With the introduction of the My Connections app, the Degree one  results provides a list of all pertinent relationships.  The Missing Links app kind of does converse (Step One results), showing what relationships are missing.  Neither of the apps label the results as Parents, Spouses and Children.  And they do not assess if they have been sourced.   

I believe we also have several apps that can assess if a birth, death, or marriage date has been provided if appropriately supplied but again, nothing that assesses if these have been sourced.

The key to GD profiles is that you have found all Degree One (or Step one) connections and have provided at least one source showing true evidence for the connection and you have established BMD Vitals with a least once source of evidence for each.

I would love to see a button on a profile that you could push that would perform the Degree one connection with relationship labels (parent, spouse, children) and with BMD dates.  

One could then assess if there are missing relationships and dates.

And from I have seen, it seems like if we included appropriate information in the Biography (with a source for each of these connections and vitals), we can then assess if they have each been properly sourced and determine what is needed of if the profile has truly been Genealogically Defined.

Again, the idea is to provide a tool to aid in researching and creating a basic or beginning level profile that could be considered completed with appropriate evidence to support all key components.

Thoughts? 

Comments

I forgot to add Siblings to the list.

Also, I meant to add that Ian's Family Group App (G2G Discussion) does a good job at this as well.

---

My thoughts are that this can be fully implemented now through the inefficient use/creation of many maintenance categories and personal "maintenance" categories. But it is doable now. Anyone can do it now if they desire too.

---

Hmm..categories and stickers completely miss my intention.  Identifying GD profiles is not the goal.  Rather I am looking for a tool that aids in the development of well documented profiles.  

Well researched, source laden profiles makes WikiTree better.  Identifying all degree one relatives with a good source should be a goal of all profiles.  GD just provides a construct for research.  The tool would aid in facilitating that research.

---

The stickers and categories are our tools that we use along the way in developing the profile.

The Sourcer browser extension is a tool that aids in facilitating research.

The "search" buttons on Ancestry and FamilySearch are tools that aid in facilitating research. WikiTree members just need to use it.

---

"I would love to see a button on a profile that you could push that would perform the Degree one connection with relationship labels (parent, spouse, children) and with BMD dates."

The Family Group app icon is available now on a profile that you can use.

Family Group app

The 'home' icon that is on a profile page comes close to displaying everything except for the marriage information.

You must be using the WikiTree BEE browser extension to see the icon.

---

[[Space:Genealogically_Defined|Genealogically Defined]] doesn't mention siblings either.

---

Thank you Jeff, my error.  I always try to include siblings in research, part of the FFAN (Family, Friends, Associates, Neighbors) research.

I suppose the reason is that siblings would be the children of the parents and would be researched under the parent's profile.

Often, the record establishing parents for a sibling can serve as evidence for one's relationship to the parents if other evidence connects the siblings.  So I am always thinking siblings when researching.

Thanks for the catch.

---

You and me both, Michael. I suspect the motivation for not including it is the ripple effect. If one sibling is problematic, not only do the parents not meet the criteria, but neither do any of the siblings. I appreciate the simple elegance of 1) each parent, 2) each spouse, 3) each child. Each genealogically defined person makes every connected parent, spouse, and child one step closer to meeting the criteria. I've only become aware of this concept because of your post. I've started with my profile and I'm working my way up my family tree to try to make each ancestor profile meet this criteria. I really appreciate your original post. Thank you.

Answer 1

The issue with deciding if something is properly sourced, is judging how reliable the source is.  I think it will be hard to programmatically determine this.

What I currently do determine if a profile is properly source is to record all the profile I'm interested in, in a spreadsheet, with columns for Birth, Death, Marriages, Spouses, Children & Occupation, and I mark them there as to completeness.

(Marriages means I have added a source for the marriage, Spouses if I've added profiles for the spouses)

Comments

Yes, human eyes need to review the evidence.  But the process could provide what source was used or if a source was missing for a specific event.

---

A Sticker could also be created to record whatever.

Example:

{{Profile Needs | birth= | death= | marriage= | spouses= | children= | occupation= | etc }}

Then an app or extension can be implemented to build a table (spreadsheet) for further analysis of the profiles.

Note:  An app or extension can be implemented now to build a table (spreadsheet) for those profiles that have made use of maintenance categories.

---

We have such categories in the Netherlands Project with the Netherlands Sticker, but instead we have needs, needs1, needs2 which can have the values Biography, Birth, LNAB, More Records, PPP, Profiles Created, which will actually create categories.

---

Coen, why are there three parameters (needs | needs1 | need2) which seem to do the same thing?

---

It has three needs so you can tag it with up to three out of the 7 categories.

---

So, if you needed the other 4 items (categories), would you just list the sticker two more times?

---

No, you don't put the sticker twice, if it is missing birth, death & marriage, then I use the {{unsourced|Netherlands}} instead, as then you don't have any sources, or you could use the generic More Records to indicate that you need more than one source.

---

You can also add them as categories. And of course you can search for them in Wikitree+. Many projects have maintenance categories which can help track what specifically is needed. 

I wish these categories existed for all countries and we had consistent definitions of how to use at least some of them so that it would help to identify where work is needed. This also seems much better than setting up my own categories for a similar purpose.

---

W, what I am trying to sort out is whether we can put something on a profile that an app can read to remind us what work is needed on that profile versus adding several categories to a profile to accomplish the same task. It should be simple enough to implement to where I don't need to keep a separate list or spreadsheet (the table that I would produce should suffice).

Personally, I would like to look at a table of profiles (i.e., my watchlist) and see what things I need to work on.

---

You could search for your wikitree+ wikitree id as manager and look for the needs categories. (The categories also display in the far right column of a search so with the needs categories you already have a table of what you manage and what needs to be done.) I would think an app could do something with these search results and display in an even better way.

Answer 2

A Genealogically Defined Sticker could be created to place on a profile as such:

{{GD Sticker | parents=yes | spouse/spouses=yes | child/children=yes }}

Then an app or extension can be implemented to build a table (spreadsheet) for further analysis of the profiles.

Comments

I like the idea of a sticker.  However, it was proposed in 2017:

https://www.wikitree.com/g2g/482334/proposal-genealogically-defined-sticker

---

So [[Space:Genealogically Defined|Genealogically Defined]] is being used instead of an "official" sticker such as {{Genealogically Defined Sticker}}.

---

We have such categories in the Netherlands Sticker, but instead we have needs, needs1, needs2 which can have the values Biography, Birth, LNAB, More Records, PPP, Profiles Created

---

Stickers are decorative elements and can be removed by any contributor to a profile. Better to use categories.

---

That Netherlands sticker also add the categories, e.g. https://www.wikitree.com/wiki/Category:Netherlands_Project_Needs_Birth

Answer 3

Does anyone know how many genealogically defined profiles have been created so far this year (2022)?

Comments

How would this be knowable?

---

Hi Chase, good to see you here.

As I am sure you know but just to answer the question.

The only way that a profile can be known if to be GD. That is, that all children, siblings, spouses, parents, and all BMD vitals are entered and sourced is for human eyes to review the information. For example, how do you automate that a woman had three husbands and that they each have valid evidence to support those relationships?

Thus the only way for a profile to be marked GD, is for someone to indicate that, yes it is complete. They actually have to look at the profile and evaluate the evidence.

I am more interested in facilitating GD research than having a profile marked as such, but I suppose if one is motivated by obtaining a sticker, so be it?

I believe Ian is working on an Events app that could pull information from a header. So it could mark that a marriage is sourced, for example and provide the actual source used. Combined with his family group app, I think it can gather and report all pertinent information.  Then a researcher would need to put eyes on the profile and evaluate its completeness and veracity before indicating that it is GD.

---

Michael - What do you mean by pull info from a header? From the fields? How would the app be able to tell if the info was sourced?

---

Right, I think our idea about a scale of source quality is the closest to this. That isn't real yet, and would still be voluntary, manual work

---

Hi Chase, I left to take my wife and daughter out for Mother's day.

Well, you have discovered a flaw in my thinking.  I know Ian was pulling sources into his Family Group report but it may not be that he could organize them by event type.

---

WikiTree BEE has a family timeline feature which includes events that have data available in either the database (through the form fields) or stickers. This includes births, marriages, deaths, and military related events. If we can support the Event template (see my recent proposal), I can also get any events that are included in the bio via the Event template. The Family Group app also presents events, but it's not easy for it to get baptism and burial data (again, see the proposal). This may explain what Michael says about an events app. With the Event template, each event has a source parameter, so, to identify genealogically defined profiles, any app could check the events templates in a bio and find if each relevant event has a source attached. It would obviously rely heavily on the Event template, but the apps wouldn't necessarily have to parse the bio each time, as WikiTree+ could store the data and make it available to apps.  

---

"any app could check the events templates in a bio and find if each relevant event has a source attached"

Per my answer to the original question, creating an app that can determine whether an event has a relevant source attached would be the hard part.

---

Hi Chase.  Sorry if I didn't explain it well.  Have you looked at my proposal about supporting the Event template?  Aleš has a made a really useful template that can produce timelines of biographies.  The more important thing from my point of view is that it can also store data in the bio that we can't store in the database due to a lack of the required database fields. The Event template has a source parameter which links a source with a specific event.  This answers your question about the app determining if there's a relevant source.  Having said that, I don't think it's impossible (though, yes, maybe very hard) to make an app that would identify all the relevant sources.  In the BEE Auto Bio, in order to create inline citations, I look for the relevant source, so it says "John Smith was born in Glasgow, Scotland on June 7, 1902, [etc.]" (for example) and then the citation for his baptism is added inline. It is possible, though not always easy, to search the bio for citations that support the data.

---

So Chase, to your point, if a user has to be the one to add the Event template in the first place, which would theoretically include the source citation, then the user would be the one evaluating whether the source was relevant. 

Trust is an issue, but if we combine that work with the Source evaluation work that has been discussed here and here, perhaps that would give us a better feeling about whether the citation is relevant without having to go do that work through AI

---

If AI was the solution, Ancestry and FamilySearch wouldn't be giving us Hints, they would be attaching the sources to the member's trees.

---

@Ian - My bad. I hadn't realized that the Event Template permits (requires?) providing a source for an event. If a profile had properly complete Event Templates for all the events needed to determine if the profile was Genealogically Defined, it would be easy for any app to determine whether the profile was Genealogically Defined. But seems like a long-shot that Event Templates will ever become widely and consistently used, so I don't see them as a practical solution.

Could you explain how WikiTree BEE Auto Bio determines if a source is related to an event?

---

It's not too complicated.  Let's say I need to find a source for a birth/baptism, for example, from the various citations on the page...  All the sources are checked for the name of the profile person and words like 'birth', 'baptism', or 'Christening'. If the date given is similar to the birth year of the profile person, then we can be pretty sure that that's the right citation.  It's far from perfect, but it mostly works.

Edit: Oh, and I agree that the Event template is not the ideal answer - that would be things like extra database fields - but if people want their extra data (like dates and locations of baptisms, burials, graduations, etc.) to be available to apps for whatever reason (like letting the Family Group app work better), some people might use it.

Answer 4

I don't think that creating an app that identifies whether particular relationships have been sourced is very doable.

1. While the profile fields identify the relationships we want to check, there is nothing in profiles that associates the fields with any sources in the profiles. I don't think there is any workaround for this.
2. For factual statements in a bio, if in-line citations are used, those citations associate sources with particular factual statements. However, the app would need to be able to read and interpret the narrative bio to identify the relationships to check. To do that, the app would need to utilize natural language processing AI. Theoretically doable, but it would depend on whether there are free 3d party NLP components available yet.

Even if we are able to accurately identify the asserted relationships we want to check and the sources (if any) provided in support of those relationships, I assume we would want to know that the source is relevant (actually supports the statement), which would require the app to be able to find and read the source. This would require that the source be available online and, once the source was found, would need to use NLP AI again to determine if the source supported asserted relationship. This sounds very hard.

 

Comments

1. While the profile fields identify the relationships we want to check, there is nothing in profiles that associates the fields with any sources in the profiles. I don't think there is any workaround for this.

The work around (solution) is the Events Template that was discussed here:

• Proposal: Support the Event template

Is it an optimal solution?:  NO

Is it a work around solution?:  YES

---

My bad. I hadn't realized that the Event Template permits (requires?) providing a source for an event. If a profile had properly complete Event Templates for all the events needed to determine if the profile was Genealogically Defined, it would be easy for any app to determine whether the profile was Genealogically Defined. But seems like a long-shot that Event Templates will ever become widely and consistently used, so I don't see them as a practical solution.