Merge fan chart

johnsullivan

I am working through multiple merges to find where an incorrect merge occurred, only to find merge upon merge upon merge. The only way I can make sense of it is to create a spreadsheet and list all the merges and the surviving entities. Even then it will take some careful analysis to identify the mistake and correct it.

I had a thought that creating a "fan chart" of the merges would be a good visual way to understand the history and look for mistakes.

What do others think? A good idea or am I just staying up too late and my mind has gone to putty?

MandyShaw1

I think that's a fantastic thought @johnsullivan, you could maybe put it on Suggest an Idea.

Alan E. Brown

I can certainly see some value in having some sort of visualization of the merges for a person profile. I'm having a hard time understanding how a fan chart would work for that, since a fan chart is very oriented towards each node having exactly two parents, and works best when most nodes have two parents. With merges, there certainly are two parents to a node (the deleted profile and the former version of the surviving profile), but in many cases only one of the nodes has two parents. And unlike a fan chart, there's no particular meaning for a "generation" of merge ancestors.

But even if a fan chart isn't the optimal visualization, it certainly seems worthwhile to explore ways to visualize the data. It could be that some other sort of tree visualization would work.

Adrian Bruce1

I'd agree with @Alan E. Brown - some sort of visualisation would seem a jolly good thing but the conventional family fan seems fraught to me.

Some sort of branching "tree" would seem the way forward - if PIDs 1, 2, 3 are all merged in that order into PID Prime, then the representation should show PID Prime going back into prehistory (i.e. 2012?) but with PIDs 1, 2 and 3 joining PID Prime in succession. Which is why a conventional fan chart doesn't work - PID Prime can't go back into previous generations in a conventional fan chart.

However, the crucial point is yes, the suggestion of some sort of graphical representation sounds excellent… Bear in mind that PIDs 1, 2, and 3 can themselves be the result of merges and this also needs to be shown…

Gordon Collett

Sounds like an interesting idea. The best structure would probably be an "inverse descendency" structure.

This is basically what I create by hand when working to untangle a complex conglomerate of incorrect merges that combined multiple different people into one profile.

Mary Anna Ebert

@johnsullivan Thanks for the suggestion!

We know untangling multiple merges is frustrating. We are investigating how to best help users undo incorrect merges, and hopefully, sometime soon, we will have an answer.

I know some of the people I've talked to, use a spreadsheet, which isn't ideal.

If you have any suggestions, we would love to hear more!

Thanks

Rhonda Budvarson

Thank you everyone for your input. I have started a separate discussion, with links to merge presentations. Please take a look and chime in with your "best practice steps" :)

Also, please share the links if you find other merge presentations that are really good.

https://community.familysearch.org/en/discussion/183186/merge-experience-presentation-thursday-13-nov-25-at-11am-mst#latest